use_case_examples/disease_prediction/HealthCarePrediction.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Health Prediction using Fully Homomorphic Encryption\n",
    "\n",
    "Processing medical data has long posed a significant challenge in the realm of machine learning primarily due to the sensitive nature of the data and the rigorous regulations safeguarding patient privacy.\n",
    "\n",
    "This notebook showcases the use of Concrete ML for a diagnosis task that relies on sensitive information: patient symptoms. Additionally, the notebook provides guidance on how to select the most appropriate Concrete model for real-life deployment in this [Hugging Face space](https://huggingface.co/spaces/zama-fhe/health_prediction)."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Import libraries"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import shutil\n",
    "from pathlib import Path\n",
    "from time import time\n",
    "from typing import Callable, Dict\n",
    "\n",
    "import matplotlib.pyplot as plt\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "from preprocessing import prepare_data\n",
    "from sklearn.model_selection import GridSearchCV\n",
    "from tqdm.notebook import tqdm\n",
    "\n",
    "from concrete.ml.sklearn import LogisticRegression as ConcreteLogisticRegression\n",
    "from concrete.ml.sklearn import XGBClassifier as ConcreteXGBClassifier"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Data-set\n",
    "\n",
    "\n",
    "The data used are available via this [link](https://github.com/anujdutt9/Disease-Prediction-from-Symptoms/tree/master/dataset).\n",
    "\n",
    "\n",
    "For the sake of simplicity, we performed a preliminary preprocessing on the data, such as correcting column names and encoding the target column that we saved in [Training_preprocessed.csv](./data/Training_preprocessed.csv) for the training set and [Testing_preprocessed.csv](./data/Testing_preprocessed.csv) for test set."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>itching</th>\n",
       "      <th>skin_rash</th>\n",
       "      <th>nodal_skin_eruptions</th>\n",
       "      <th>continuous_sneezing</th>\n",
       "      <th>shivering</th>\n",
       "      <th>chills</th>\n",
       "      <th>joint_pain</th>\n",
       "      <th>stomach_pain</th>\n",
       "      <th>acidity</th>\n",
       "      <th>ulcers_on_tongue</th>\n",
       "      <th>...</th>\n",
       "      <th>scurving</th>\n",
       "      <th>skin_peeling</th>\n",
       "      <th>silver_like_dusting</th>\n",
       "      <th>small_dents_in_nails</th>\n",
       "      <th>inflammatory_nails</th>\n",
       "      <th>blister</th>\n",
       "      <th>red_sore_around_nose</th>\n",
       "      <th>yellow_crust_ooze</th>\n",
       "      <th>prognosis</th>\n",
       "      <th>prognosis_encoded</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>Fungal Infection</td>\n",
       "      <td>14</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>Fungal Infection</td>\n",
       "      <td>14</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>Fungal Infection</td>\n",
       "      <td>14</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>Fungal Infection</td>\n",
       "      <td>14</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>Fungal Infection</td>\n",
       "      <td>14</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 130 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   itching  skin_rash  nodal_skin_eruptions  continuous_sneezing  shivering  \\\n",
       "0      1.0        1.0                   1.0                  0.0        0.0   \n",
       "1      0.0        1.0                   1.0                  0.0        0.0   \n",
       "2      1.0        0.0                   1.0                  0.0        0.0   \n",
       "3      1.0        1.0                   0.0                  0.0        0.0   \n",
       "4      1.0        1.0                   1.0                  0.0        0.0   \n",
       "\n",
       "   chills  joint_pain  stomach_pain  acidity  ulcers_on_tongue  ...  scurving  \\\n",
       "0     0.0         0.0           0.0      0.0               0.0  ...       0.0   \n",
       "1     0.0         0.0           0.0      0.0               0.0  ...       0.0   \n",
       "2     0.0         0.0           0.0      0.0               0.0  ...       0.0   \n",
       "3     0.0         0.0           0.0      0.0               0.0  ...       0.0   \n",
       "4     0.0         0.0           0.0      0.0               0.0  ...       0.0   \n",
       "\n",
       "   skin_peeling  silver_like_dusting  small_dents_in_nails  \\\n",
       "0           0.0                  0.0                   0.0   \n",
       "1           0.0                  0.0                   0.0   \n",
       "2           0.0                  0.0                   0.0   \n",
       "3           0.0                  0.0                   0.0   \n",
       "4           0.0                  0.0                   0.0   \n",
       "\n",
       "   inflammatory_nails  blister  red_sore_around_nose  yellow_crust_ooze  \\\n",
       "0                 0.0      0.0                   0.0                0.0   \n",
       "1                 0.0      0.0                   0.0                0.0   \n",
       "2                 0.0      0.0                   0.0                0.0   \n",
       "3                 0.0      0.0                   0.0                0.0   \n",
       "4                 0.0      0.0                   0.0                0.0   \n",
       "\n",
       "          prognosis  prognosis_encoded  \n",
       "0  Fungal Infection                 14  \n",
       "1  Fungal Infection                 14  \n",
       "2  Fungal Infection                 14  \n",
       "3  Fungal Infection                 14  \n",
       "4  Fungal Infection                 14  \n",
       "\n",
       "[5 rows x 130 columns]"
      ]
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Load the train and testing sets\n",
    "\n",
    "df_train, df_test = prepare_data()\n",
    "\n",
    "TARGET_COLUMN = [\"prognosis_encoded\", \"prognosis\"]\n",
    "\n",
    "y_train = df_train[TARGET_COLUMN[0]].values.flatten()\n",
    "y_test = df_test[TARGET_COLUMN[0]].values.flatten()\n",
    "\n",
    "X_train = df_train.drop(TARGET_COLUMN, axis=1)\n",
    "X_test = df_test.drop(TARGET_COLUMN, axis=1)\n",
    "\n",
    "df_train.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "The dataset contains: 128 distinct symptoms and 41 unique diseases.\n"
     ]
    }
   ],
   "source": [
    "print(\n",
    "    f\"The dataset contains: {X_train.shape[1]} distinct symptoms and \"\n",
    "    f\"{len(np.unique(y_train))} unique diseases.\"\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1500x500 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# The Training data-set is well-balanced\n",
    "\n",
    "value_counts = df_train[TARGET_COLUMN[1]].value_counts().sort_values(ascending=False)\n",
    "\n",
    "plt.figure(figsize=(15, 5))\n",
    "\n",
    "ax = value_counts.plot.bar(color=\"orange\")\n",
    "ax.set_xlabel(\"Diseases\")\n",
    "ax.set_ylim(0, 150)\n",
    "ax.set_ylabel(\"Value Count\")\n",
    "ax.set_title(\"Disease Frequency\")\n",
    "\n",
    "plt.xticks(rotation=80)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Accuracy vs speed trade-off for FHE ML models\n",
    "\n",
    "FHE programs are slower due to various factors, including the high cost of PBS operations and current hardware limitations. However, ongoing research is actively improving the performance of FHE to overcome these constraints. To mitigate these constraints in real-life use-cases, developers must search for an optimal model configuration that obtains an interesting trade-off between speed and accuracy. To do so, we use grid search, evaluating the model using FHE simulation mode."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "def grid_search(grid_param: Dict, clf: Callable) -> GridSearchCV:\n",
    "    \"\"\"\n",
    "    Perform a grid search to find the best hyper-parameters for a given classifier.\n",
    "\n",
    "    Args:\n",
    "        grid_param (Dict): The hyper-parameters to be tuned\n",
    "        clf (Callable): The given classifier\n",
    "\n",
    "    Returns:\n",
    "        GridSearchCV: The fitted GridSearchCV object.\n",
    "\n",
    "    \"\"\"\n",
    "    grid_search = GridSearchCV(\n",
    "        clf,\n",
    "        grid_param,\n",
    "        cv=4,\n",
    "        verbose=1,\n",
    "        n_jobs=-1,\n",
    "    ).fit(X_train, y_train)\n",
    "\n",
    "    # The best model\n",
    "    print(f\"Best parameters: {grid_search.best_params_}\")\n",
    "    print(f\"Best score: {grid_search.best_score_:.2%}\")\n",
    "\n",
    "    return grid_search"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "def fhe_evaluation(\n",
    "    grid_search: GridSearchCV, clf_class: Callable, grid_param: Dict, X_test: np.ndarray\n",
    ") -> pd.DataFrame:\n",
    "    \"\"\"\n",
    "    Mesuare the time inference in FHE for each possible combinaison.\n",
    "\n",
    "    Args:\n",
    "        grid_search (GridSearchCV): Grid search object used to find the best hyperparameters\n",
    "        clf_class (Callable): The classifier model to be evaluated\n",
    "        grid_param (Dict): The hyper-parameters associated to the givin classifier\n",
    "        X_test (np.ndarray): One sample used to measure the time inference in FHE\n",
    "\n",
    "    Returns:\n",
    "        pd.DataFrame: All combinaison of hyper-parameters and their mean test scores, standard test\n",
    "            scores, and time inference in FHE\n",
    "\n",
    "    \"\"\"\n",
    "    # Save the grid search results as a DataFrame\n",
    "    grid_results = pd.DataFrame(grid_search.cv_results_)\n",
    "    grid_results.columns = grid_results.columns.str.replace(\"param_\", \"\")\n",
    "    # Pick up relevent columns\n",
    "    grid_results = grid_results[list(grid_param.keys()) + [\"mean_test_score\", \"std_test_score\"]]\n",
    "    # Sort the results by mean_test_score in descending order\n",
    "    grid_results = grid_results.sort_values(by=\"mean_test_score\", ascending=False)\n",
    "\n",
    "    fhe_inference_time = []\n",
    "    for _, param in tqdm(grid_results[list(grid_param.keys())].iterrows()):\n",
    "        # Instanciation\n",
    "        clf = clf_class(**dict(param))\n",
    "        # Training\n",
    "        clf.fit(X_train, y_train)\n",
    "        fhe_circuit = clf.compile(X_train)\n",
    "        # Key generation\n",
    "        fhe_circuit.client.keygen(force=False)\n",
    "        # Inference\n",
    "        start_time = time()\n",
    "        _ = clf.predict(X_test, fhe=\"execute\")\n",
    "        fhe_inference_time.append(time() - start_time)\n",
    "    grid_results[\"fhe_inference_time\"] = fhe_inference_time\n",
    "\n",
    "    return grid_results"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We start with XGB classifier:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fitting 4 folds for each of 36 candidates, totalling 144 fits\n",
      "Best parameters: {'max_depth': 3, 'n_bits': 7, 'n_estimators': 2}\n",
      "Best score: 99.88%\n"
     ]
    }
   ],
   "source": [
    "grid_xgb_param = {\n",
    "    \"n_bits\": [2, 3, 4, 5, 6, 7],\n",
    "    \"max_depth\": [1, 2, 3],\n",
    "    \"n_estimators\": [2, 3],\n",
    "}\n",
    "\n",
    "gr_xgb = grid_search(grid_xgb_param, ConcreteXGBClassifier())"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Then, we test the Logistic Regression model:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fitting 4 folds for each of 64 candidates, totalling 256 fits\n",
      "Best parameters: {'C': 0.5, 'multi_class': 'auto', 'n_bits': 7, 'solver': 'newton-cg'}\n",
      "Best score: 100.00%\n"
     ]
    }
   ],
   "source": [
    "grid_logit_param = {\n",
    "    \"C\": [0.001, 0.5, 0.9, 1.0],\n",
    "    \"n_bits\": [7, 10, 13, 14],\n",
    "    \"solver\": [\"newton-cg\", \"sag\", \"saga\", \"lbfgs\"],\n",
    "    \"multi_class\": [\"auto\"],\n",
    "}\n",
    "\n",
    "gs_logit = grid_search(grid_logit_param, ConcreteLogisticRegression())"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The best model configuration may not necessarily mean the best performing model in the FHE realm. To highlight this nuance, we measure the inference timing for a single random sample in real FHE mode for each configuration."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Random disease: Migraine\n"
     ]
    }
   ],
   "source": [
    "# Take a random input as a test case to measure the execution time of all the Concrete ML models\n",
    "random_row = df_test.sample().values[0]\n",
    "\n",
    "random_x, random_y = random_row[:-2][np.newaxis, :], random_row[-2]\n",
    "\n",
    "print(f\"Random disease: {random_y}\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "5d6e30e42a0f48539d15a144a66a7c65",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "0it [00:00, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "xgb_results = fhe_evaluation(gr_xgb, ConcreteXGBClassifier, grid_xgb_param, random_x)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "7d1c65d3383e489dbd3c34716267ebc2",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "0it [00:00, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "logit_results = fhe_evaluation(gs_logit, ConcreteLogisticRegression, grid_logit_param, random_x)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "For each Concrete ML model, we display its score along with the corresponding inference timing.\n",
    "This enables us to evaluate the model's effectiveness in terms of performance and its computational efficiency."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<style type=\"text/css\">\n",
       "#T_a7ce0_row0_col0, #T_a7ce0_row0_col3, #T_a7ce0_row1_col0, #T_a7ce0_row1_col5, #T_a7ce0_row4_col0, #T_a7ce0_row5_col0, #T_a7ce0_row14_col0, #T_a7ce0_row17_col0 {\n",
       "  background-color: #023858;\n",
       "  color: #f1f1f1;\n",
       "}\n",
       "#T_a7ce0_row0_col5, #T_a7ce0_row32_col3, #T_a7ce0_row33_col3 {\n",
       "  background-color: #2a88bc;\n",
       "  color: #f1f1f1;\n",
       "}\n",
       "#T_a7ce0_row1_col3 {\n",
       "  background-color: #02395a;\n",
       "  color: #f1f1f1;\n",
       "}\n",
       "#T_a7ce0_row2_col0, #T_a7ce0_row3_col0, #T_a7ce0_row6_col0, #T_a7ce0_row7_col0, #T_a7ce0_row12_col0, #T_a7ce0_row13_col0 {\n",
       "  background-color: #0567a2;\n",
       "  color: #f1f1f1;\n",
       "}\n",
       "#T_a7ce0_row2_col3, #T_a7ce0_row3_col3 {\n",
       "  background-color: #023c5f;\n",
       "  color: #f1f1f1;\n",
       "}\n",
       "#T_a7ce0_row2_col5 {\n",
       "  background-color: #a5bddb;\n",
       "  color: #000000;\n",
       "}\n",
       "#T_a7ce0_row3_col5 {\n",
       "  background-color: #d5d5e8;\n",
       "  color: #000000;\n",
       "}\n",
       "#T_a7ce0_row4_col3, #T_a7ce0_row5_col3 {\n",
       "  background-color: #023f64;\n",
       "  color: #f1f1f1;\n",
       "}\n",
       "#T_a7ce0_row4_col5 {\n",
       "  background-color: #0a73b2;\n",
       "  color: #f1f1f1;\n",
       "}\n",
       "#T_a7ce0_row5_col5 {\n",
       "  background-color: #79abd0;\n",
       "  color: #f1f1f1;\n",
       "}\n",
       "#T_a7ce0_row6_col3, #T_a7ce0_row7_col3 {\n",
       "  background-color: #034369;\n",
       "  color: #f1f1f1;\n",
       "}\n",
       "#T_a7ce0_row6_col5 {\n",
       "  background-color: #c8cde4;\n",
       "  color: #000000;\n",
       "}\n",
       "#T_a7ce0_row7_col5 {\n",
       "  background-color: #e0dded;\n",
       "  color: #000000;\n",
       "}\n",
       "#T_a7ce0_row8_col0, #T_a7ce0_row9_col0, #T_a7ce0_row10_col0, #T_a7ce0_row11_col0, #T_a7ce0_row15_col0, #T_a7ce0_row16_col0 {\n",
       "  background-color: #4295c3;\n",
       "  color: #f1f1f1;\n",
       "}\n",
       "#T_a7ce0_row8_col3, #T_a7ce0_row9_col3 {\n",
       "  background-color: #034b76;\n",
       "  color: #f1f1f1;\n",
       "}\n",
       "#T_a7ce0_row8_col5 {\n",
       "  background-color: #eae6f1;\n",
       "  color: #000000;\n",
       "}\n",
       "#T_a7ce0_row9_col5, #T_a7ce0_row10_col5 {\n",
       "  background-color: #f4eef6;\n",
       "  color: #000000;\n",
       "}\n",
       "#T_a7ce0_row10_col3 {\n",
       "  background-color: #034e7b;\n",
       "  color: #f1f1f1;\n",
       "}\n",
       "#T_a7ce0_row11_col3 {\n",
       "  background-color: #034f7d;\n",
       "  color: #f1f1f1;\n",
       "}\n",
       "#T_a7ce0_row11_col5, #T_a7ce0_row28_col5 {\n",
       "  background-color: #f9f2f8;\n",
       "  color: #000000;\n",
       "}\n",
       "#T_a7ce0_row12_col3, #T_a7ce0_row13_col3 {\n",
       "  background-color: #045280;\n",
       "  color: #f1f1f1;\n",
       "}\n",
       "#T_a7ce0_row12_col5 {\n",
       "  background-color: #f1ebf4;\n",
       "  color: #000000;\n",
       "}\n",
       "#T_a7ce0_row13_col5 {\n",
       "  background-color: #e7e3f0;\n",
       "  color: #000000;\n",
       "}\n",
       "#T_a7ce0_row14_col3, #T_a7ce0_row15_col3, #T_a7ce0_row16_col3, #T_a7ce0_row17_col3, #T_a7ce0_row18_col3 {\n",
       "  background-color: #045382;\n",
       "  color: #f1f1f1;\n",
       "}\n",
       "#T_a7ce0_row14_col5 {\n",
       "  background-color: #93b5d6;\n",
       "  color: #000000;\n",
       "}\n",
       "#T_a7ce0_row15_col5, #T_a7ce0_row24_col5, #T_a7ce0_row27_col5, #T_a7ce0_row29_col5, #T_a7ce0_row31_col5 {\n",
       "  background-color: #fdf5fa;\n",
       "  color: #000000;\n",
       "}\n",
       "#T_a7ce0_row16_col5 {\n",
       "  background-color: #faf3f9;\n",
       "  color: #000000;\n",
       "}\n",
       "#T_a7ce0_row17_col5 {\n",
       "  background-color: #c4cbe3;\n",
       "  color: #000000;\n",
       "}\n",
       "#T_a7ce0_row18_col0, #T_a7ce0_row19_col0, #T_a7ce0_row21_col0, #T_a7ce0_row24_col0, #T_a7ce0_row25_col0, #T_a7ce0_row27_col0 {\n",
       "  background-color: #9cb9d9;\n",
       "  color: #000000;\n",
       "}\n",
       "#T_a7ce0_row18_col5 {\n",
       "  background-color: #fcf4fa;\n",
       "  color: #000000;\n",
       "}\n",
       "#T_a7ce0_row19_col3 {\n",
       "  background-color: #045585;\n",
       "  color: #f1f1f1;\n",
       "}\n",
       "#T_a7ce0_row19_col5 {\n",
       "  background-color: #fef6fb;\n",
       "  color: #000000;\n",
       "}\n",
       "#T_a7ce0_row20_col0, #T_a7ce0_row22_col0, #T_a7ce0_row23_col0, #T_a7ce0_row28_col0, #T_a7ce0_row29_col0 {\n",
       "  background-color: #dbdaeb;\n",
       "  color: #000000;\n",
       "}\n",
       "#T_a7ce0_row20_col3 {\n",
       "  background-color: #045687;\n",
       "  color: #f1f1f1;\n",
       "}\n",
       "#T_a7ce0_row20_col5, #T_a7ce0_row25_col5 {\n",
       "  background-color: #faf2f8;\n",
       "  color: #000000;\n",
       "}\n",
       "#T_a7ce0_row21_col3, #T_a7ce0_row22_col3, #T_a7ce0_row23_col3, #T_a7ce0_row24_col3, #T_a7ce0_row25_col3 {\n",
       "  background-color: #04588a;\n",
       "  color: #f1f1f1;\n",
       "}\n",
       "#T_a7ce0_row21_col5 {\n",
       "  background-color: #f8f1f8;\n",
       "  color: #000000;\n",
       "}\n",
       "#T_a7ce0_row22_col5, #T_a7ce0_row23_col5 {\n",
       "  background-color: #f5eff6;\n",
       "  color: #000000;\n",
       "}\n",
       "#T_a7ce0_row26_col0 {\n",
       "  background-color: #dbdaeb;\n",
       "  color: #000000;\n",
       "  background-color: green;\n",
       "}\n",
       "#T_a7ce0_row26_col1, #T_a7ce0_row26_col2, #T_a7ce0_row26_col4 {\n",
       "  background-color: green;\n",
       "}\n",
       "#T_a7ce0_row26_col3 {\n",
       "  background-color: #045a8d;\n",
       "  color: #f1f1f1;\n",
       "  background-color: green;\n",
       "}\n",
       "#T_a7ce0_row26_col5 {\n",
       "  background-color: #fff7fb;\n",
       "  color: #000000;\n",
       "  background-color: green;\n",
       "}\n",
       "#T_a7ce0_row27_col3, #T_a7ce0_row28_col3, #T_a7ce0_row29_col3 {\n",
       "  background-color: #045a8d;\n",
       "  color: #f1f1f1;\n",
       "}\n",
       "#T_a7ce0_row30_col0, #T_a7ce0_row31_col0, #T_a7ce0_row32_col0, #T_a7ce0_row33_col0, #T_a7ce0_row34_col0, #T_a7ce0_row34_col3, #T_a7ce0_row35_col0, #T_a7ce0_row35_col3, #T_a7ce0_row35_col5 {\n",
       "  background-color: #fff7fb;\n",
       "  color: #000000;\n",
       "}\n",
       "#T_a7ce0_row30_col3, #T_a7ce0_row31_col3 {\n",
       "  background-color: #056dac;\n",
       "  color: #f1f1f1;\n",
       "}\n",
       "#T_a7ce0_row30_col5 {\n",
       "  background-color: #e9e5f1;\n",
       "  color: #000000;\n",
       "}\n",
       "#T_a7ce0_row32_col5, #T_a7ce0_row34_col5 {\n",
       "  background-color: #fef6fa;\n",
       "  color: #000000;\n",
       "}\n",
       "#T_a7ce0_row33_col5 {\n",
       "  background-color: #fbf4f9;\n",
       "  color: #000000;\n",
       "}\n",
       "</style>\n",
       "<table id=\"T_a7ce0\">\n",
       "  <thead>\n",
       "    <tr>\n",
       "      <th class=\"blank level0\" >&nbsp;</th>\n",
       "      <th id=\"T_a7ce0_level0_col0\" class=\"col_heading level0 col0\" >n_bits</th>\n",
       "      <th id=\"T_a7ce0_level0_col1\" class=\"col_heading level0 col1\" >max_depth</th>\n",
       "      <th id=\"T_a7ce0_level0_col2\" class=\"col_heading level0 col2\" >n_estimators</th>\n",
       "      <th id=\"T_a7ce0_level0_col3\" class=\"col_heading level0 col3\" >mean_test_score</th>\n",
       "      <th id=\"T_a7ce0_level0_col4\" class=\"col_heading level0 col4\" >std_test_score</th>\n",
       "      <th id=\"T_a7ce0_level0_col5\" class=\"col_heading level0 col5\" >fhe_inference_time</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th id=\"T_a7ce0_level0_row0\" class=\"row_heading level0 row0\" >34</th>\n",
       "      <td id=\"T_a7ce0_row0_col0\" class=\"data row0 col0\" >7</td>\n",
       "      <td id=\"T_a7ce0_row0_col1\" class=\"data row0 col1\" >3</td>\n",
       "      <td id=\"T_a7ce0_row0_col2\" class=\"data row0 col2\" >2</td>\n",
       "      <td id=\"T_a7ce0_row0_col3\" class=\"data row0 col3\" >0.998780</td>\n",
       "      <td id=\"T_a7ce0_row0_col4\" class=\"data row0 col4\" >0.001220</td>\n",
       "      <td id=\"T_a7ce0_row0_col5\" class=\"data row0 col5\" >50.848390</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_a7ce0_level0_row1\" class=\"row_heading level0 row1\" >35</th>\n",
       "      <td id=\"T_a7ce0_row1_col0\" class=\"data row1 col0\" >7</td>\n",
       "      <td id=\"T_a7ce0_row1_col1\" class=\"data row1 col1\" >3</td>\n",
       "      <td id=\"T_a7ce0_row1_col2\" class=\"data row1 col2\" >3</td>\n",
       "      <td id=\"T_a7ce0_row1_col3\" class=\"data row1 col3\" >0.997561</td>\n",
       "      <td id=\"T_a7ce0_row1_col4\" class=\"data row1 col4\" >0.002439</td>\n",
       "      <td id=\"T_a7ce0_row1_col5\" class=\"data row1 col5\" >77.306330</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_a7ce0_level0_row2\" class=\"row_heading level0 row2\" >33</th>\n",
       "      <td id=\"T_a7ce0_row2_col0\" class=\"data row2 col0\" >6</td>\n",
       "      <td id=\"T_a7ce0_row2_col1\" class=\"data row2 col1\" >3</td>\n",
       "      <td id=\"T_a7ce0_row2_col2\" class=\"data row2 col2\" >3</td>\n",
       "      <td id=\"T_a7ce0_row2_col3\" class=\"data row2 col3\" >0.995122</td>\n",
       "      <td id=\"T_a7ce0_row2_col4\" class=\"data row2 col4\" >0.004878</td>\n",
       "      <td id=\"T_a7ce0_row2_col5\" class=\"data row2 col5\" >29.819495</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_a7ce0_level0_row3\" class=\"row_heading level0 row3\" >32</th>\n",
       "      <td id=\"T_a7ce0_row3_col0\" class=\"data row3 col0\" >6</td>\n",
       "      <td id=\"T_a7ce0_row3_col1\" class=\"data row3 col1\" >3</td>\n",
       "      <td id=\"T_a7ce0_row3_col2\" class=\"data row3 col2\" >2</td>\n",
       "      <td id=\"T_a7ce0_row3_col3\" class=\"data row3 col3\" >0.995122</td>\n",
       "      <td id=\"T_a7ce0_row3_col4\" class=\"data row3 col4\" >0.004878</td>\n",
       "      <td id=\"T_a7ce0_row3_col5\" class=\"data row3 col5\" >18.353732</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_a7ce0_level0_row4\" class=\"row_heading level0 row4\" >23</th>\n",
       "      <td id=\"T_a7ce0_row4_col0\" class=\"data row4 col0\" >7</td>\n",
       "      <td id=\"T_a7ce0_row4_col1\" class=\"data row4 col1\" >2</td>\n",
       "      <td id=\"T_a7ce0_row4_col2\" class=\"data row4 col2\" >3</td>\n",
       "      <td id=\"T_a7ce0_row4_col3\" class=\"data row4 col3\" >0.992683</td>\n",
       "      <td id=\"T_a7ce0_row4_col4\" class=\"data row4 col4\" >0.007317</td>\n",
       "      <td id=\"T_a7ce0_row4_col5\" class=\"data row4 col5\" >57.148737</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_a7ce0_level0_row5\" class=\"row_heading level0 row5\" >22</th>\n",
       "      <td id=\"T_a7ce0_row5_col0\" class=\"data row5 col0\" >7</td>\n",
       "      <td id=\"T_a7ce0_row5_col1\" class=\"data row5 col1\" >2</td>\n",
       "      <td id=\"T_a7ce0_row5_col2\" class=\"data row5 col2\" >2</td>\n",
       "      <td id=\"T_a7ce0_row5_col3\" class=\"data row5 col3\" >0.992683</td>\n",
       "      <td id=\"T_a7ce0_row5_col4\" class=\"data row5 col4\" >0.007317</td>\n",
       "      <td id=\"T_a7ce0_row5_col5\" class=\"data row5 col5\" >38.101938</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_a7ce0_level0_row6\" class=\"row_heading level0 row6\" >21</th>\n",
       "      <td id=\"T_a7ce0_row6_col0\" class=\"data row6 col0\" >6</td>\n",
       "      <td id=\"T_a7ce0_row6_col1\" class=\"data row6 col1\" >2</td>\n",
       "      <td id=\"T_a7ce0_row6_col2\" class=\"data row6 col2\" >3</td>\n",
       "      <td id=\"T_a7ce0_row6_col3\" class=\"data row6 col3\" >0.990244</td>\n",
       "      <td id=\"T_a7ce0_row6_col4\" class=\"data row6 col4\" >0.009756</td>\n",
       "      <td id=\"T_a7ce0_row6_col5\" class=\"data row6 col5\" >21.930150</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_a7ce0_level0_row7\" class=\"row_heading level0 row7\" >20</th>\n",
       "      <td id=\"T_a7ce0_row7_col0\" class=\"data row7 col0\" >6</td>\n",
       "      <td id=\"T_a7ce0_row7_col1\" class=\"data row7 col1\" >2</td>\n",
       "      <td id=\"T_a7ce0_row7_col2\" class=\"data row7 col2\" >2</td>\n",
       "      <td id=\"T_a7ce0_row7_col3\" class=\"data row7 col3\" >0.990244</td>\n",
       "      <td id=\"T_a7ce0_row7_col4\" class=\"data row7 col4\" >0.009756</td>\n",
       "      <td id=\"T_a7ce0_row7_col5\" class=\"data row7 col5\" >14.774465</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_a7ce0_level0_row8\" class=\"row_heading level0 row8\" >31</th>\n",
       "      <td id=\"T_a7ce0_row8_col0\" class=\"data row8 col0\" >5</td>\n",
       "      <td id=\"T_a7ce0_row8_col1\" class=\"data row8 col1\" >3</td>\n",
       "      <td id=\"T_a7ce0_row8_col2\" class=\"data row8 col2\" >3</td>\n",
       "      <td id=\"T_a7ce0_row8_col3\" class=\"data row8 col3\" >0.984146</td>\n",
       "      <td id=\"T_a7ce0_row8_col4\" class=\"data row8 col4\" >0.015854</td>\n",
       "      <td id=\"T_a7ce0_row8_col5\" class=\"data row8 col5\" >11.284927</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_a7ce0_level0_row9\" class=\"row_heading level0 row9\" >30</th>\n",
       "      <td id=\"T_a7ce0_row9_col0\" class=\"data row9 col0\" >5</td>\n",
       "      <td id=\"T_a7ce0_row9_col1\" class=\"data row9 col1\" >3</td>\n",
       "      <td id=\"T_a7ce0_row9_col2\" class=\"data row9 col2\" >2</td>\n",
       "      <td id=\"T_a7ce0_row9_col3\" class=\"data row9 col3\" >0.984146</td>\n",
       "      <td id=\"T_a7ce0_row9_col4\" class=\"data row9 col4\" >0.015854</td>\n",
       "      <td id=\"T_a7ce0_row9_col5\" class=\"data row9 col5\" >6.420574</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_a7ce0_level0_row10\" class=\"row_heading level0 row10\" >19</th>\n",
       "      <td id=\"T_a7ce0_row10_col0\" class=\"data row10 col0\" >5</td>\n",
       "      <td id=\"T_a7ce0_row10_col1\" class=\"data row10 col1\" >2</td>\n",
       "      <td id=\"T_a7ce0_row10_col2\" class=\"data row10 col2\" >3</td>\n",
       "      <td id=\"T_a7ce0_row10_col3\" class=\"data row10 col3\" >0.981707</td>\n",
       "      <td id=\"T_a7ce0_row10_col4\" class=\"data row10 col4\" >0.018293</td>\n",
       "      <td id=\"T_a7ce0_row10_col5\" class=\"data row10 col5\" >6.404347</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_a7ce0_level0_row11\" class=\"row_heading level0 row11\" >18</th>\n",
       "      <td id=\"T_a7ce0_row11_col0\" class=\"data row11 col0\" >5</td>\n",
       "      <td id=\"T_a7ce0_row11_col1\" class=\"data row11 col1\" >2</td>\n",
       "      <td id=\"T_a7ce0_row11_col2\" class=\"data row11 col2\" >2</td>\n",
       "      <td id=\"T_a7ce0_row11_col3\" class=\"data row11 col3\" >0.980488</td>\n",
       "      <td id=\"T_a7ce0_row11_col4\" class=\"data row11 col4\" >0.019512</td>\n",
       "      <td id=\"T_a7ce0_row11_col5\" class=\"data row11 col5\" >4.192919</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_a7ce0_level0_row12\" class=\"row_heading level0 row12\" >8</th>\n",
       "      <td id=\"T_a7ce0_row12_col0\" class=\"data row12 col0\" >6</td>\n",
       "      <td id=\"T_a7ce0_row12_col1\" class=\"data row12 col1\" >1</td>\n",
       "      <td id=\"T_a7ce0_row12_col2\" class=\"data row12 col2\" >2</td>\n",
       "      <td id=\"T_a7ce0_row12_col3\" class=\"data row12 col3\" >0.979268</td>\n",
       "      <td id=\"T_a7ce0_row12_col4\" class=\"data row12 col4\" >0.020732</td>\n",
       "      <td id=\"T_a7ce0_row12_col5\" class=\"data row12 col5\" >8.230483</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_a7ce0_level0_row13\" class=\"row_heading level0 row13\" >9</th>\n",
       "      <td id=\"T_a7ce0_row13_col0\" class=\"data row13 col0\" >6</td>\n",
       "      <td id=\"T_a7ce0_row13_col1\" class=\"data row13 col1\" >1</td>\n",
       "      <td id=\"T_a7ce0_row13_col2\" class=\"data row13 col2\" >3</td>\n",
       "      <td id=\"T_a7ce0_row13_col3\" class=\"data row13 col3\" >0.979268</td>\n",
       "      <td id=\"T_a7ce0_row13_col4\" class=\"data row13 col4\" >0.020732</td>\n",
       "      <td id=\"T_a7ce0_row13_col5\" class=\"data row13 col5\" >12.459023</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_a7ce0_level0_row14\" class=\"row_heading level0 row14\" >11</th>\n",
       "      <td id=\"T_a7ce0_row14_col0\" class=\"data row14 col0\" >7</td>\n",
       "      <td id=\"T_a7ce0_row14_col1\" class=\"data row14 col1\" >1</td>\n",
       "      <td id=\"T_a7ce0_row14_col2\" class=\"data row14 col2\" >3</td>\n",
       "      <td id=\"T_a7ce0_row14_col3\" class=\"data row14 col3\" >0.978049</td>\n",
       "      <td id=\"T_a7ce0_row14_col4\" class=\"data row14 col4\" >0.021951</td>\n",
       "      <td id=\"T_a7ce0_row14_col5\" class=\"data row14 col5\" >33.308340</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_a7ce0_level0_row15\" class=\"row_heading level0 row15\" >6</th>\n",
       "      <td id=\"T_a7ce0_row15_col0\" class=\"data row15 col0\" >5</td>\n",
       "      <td id=\"T_a7ce0_row15_col1\" class=\"data row15 col1\" >1</td>\n",
       "      <td id=\"T_a7ce0_row15_col2\" class=\"data row15 col2\" >2</td>\n",
       "      <td id=\"T_a7ce0_row15_col3\" class=\"data row15 col3\" >0.978049</td>\n",
       "      <td id=\"T_a7ce0_row15_col4\" class=\"data row15 col4\" >0.021951</td>\n",
       "      <td id=\"T_a7ce0_row15_col5\" class=\"data row15 col5\" >2.147570</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_a7ce0_level0_row16\" class=\"row_heading level0 row16\" >7</th>\n",
       "      <td id=\"T_a7ce0_row16_col0\" class=\"data row16 col0\" >5</td>\n",
       "      <td id=\"T_a7ce0_row16_col1\" class=\"data row16 col1\" >1</td>\n",
       "      <td id=\"T_a7ce0_row16_col2\" class=\"data row16 col2\" >3</td>\n",
       "      <td id=\"T_a7ce0_row16_col3\" class=\"data row16 col3\" >0.978049</td>\n",
       "      <td id=\"T_a7ce0_row16_col4\" class=\"data row16 col4\" >0.021951</td>\n",
       "      <td id=\"T_a7ce0_row16_col5\" class=\"data row16 col5\" >3.349917</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_a7ce0_level0_row17\" class=\"row_heading level0 row17\" >10</th>\n",
       "      <td id=\"T_a7ce0_row17_col0\" class=\"data row17 col0\" >7</td>\n",
       "      <td id=\"T_a7ce0_row17_col1\" class=\"data row17 col1\" >1</td>\n",
       "      <td id=\"T_a7ce0_row17_col2\" class=\"data row17 col2\" >2</td>\n",
       "      <td id=\"T_a7ce0_row17_col3\" class=\"data row17 col3\" >0.978049</td>\n",
       "      <td id=\"T_a7ce0_row17_col4\" class=\"data row17 col4\" >0.021951</td>\n",
       "      <td id=\"T_a7ce0_row17_col5\" class=\"data row17 col5\" >22.814555</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_a7ce0_level0_row18\" class=\"row_heading level0 row18\" >5</th>\n",
       "      <td id=\"T_a7ce0_row18_col0\" class=\"data row18 col0\" >4</td>\n",
       "      <td id=\"T_a7ce0_row18_col1\" class=\"data row18 col1\" >1</td>\n",
       "      <td id=\"T_a7ce0_row18_col2\" class=\"data row18 col2\" >3</td>\n",
       "      <td id=\"T_a7ce0_row18_col3\" class=\"data row18 col3\" >0.978049</td>\n",
       "      <td id=\"T_a7ce0_row18_col4\" class=\"data row18 col4\" >0.021951</td>\n",
       "      <td id=\"T_a7ce0_row18_col5\" class=\"data row18 col5\" >2.558438</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_a7ce0_level0_row19\" class=\"row_heading level0 row19\" >4</th>\n",
       "      <td id=\"T_a7ce0_row19_col0\" class=\"data row19 col0\" >4</td>\n",
       "      <td id=\"T_a7ce0_row19_col1\" class=\"data row19 col1\" >1</td>\n",
       "      <td id=\"T_a7ce0_row19_col2\" class=\"data row19 col2\" >2</td>\n",
       "      <td id=\"T_a7ce0_row19_col3\" class=\"data row19 col3\" >0.976829</td>\n",
       "      <td id=\"T_a7ce0_row19_col4\" class=\"data row19 col4\" >0.023171</td>\n",
       "      <td id=\"T_a7ce0_row19_col5\" class=\"data row19 col5\" >1.404678</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_a7ce0_level0_row20\" class=\"row_heading level0 row20\" >3</th>\n",
       "      <td id=\"T_a7ce0_row20_col0\" class=\"data row20 col0\" >3</td>\n",
       "      <td id=\"T_a7ce0_row20_col1\" class=\"data row20 col1\" >1</td>\n",
       "      <td id=\"T_a7ce0_row20_col2\" class=\"data row20 col2\" >3</td>\n",
       "      <td id=\"T_a7ce0_row20_col3\" class=\"data row20 col3\" >0.975610</td>\n",
       "      <td id=\"T_a7ce0_row20_col4\" class=\"data row20 col4\" >0.024390</td>\n",
       "      <td id=\"T_a7ce0_row20_col5\" class=\"data row20 col5\" >3.628071</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_a7ce0_level0_row21\" class=\"row_heading level0 row21\" >17</th>\n",
       "      <td id=\"T_a7ce0_row21_col0\" class=\"data row21 col0\" >4</td>\n",
       "      <td id=\"T_a7ce0_row21_col1\" class=\"data row21 col1\" >2</td>\n",
       "      <td id=\"T_a7ce0_row21_col2\" class=\"data row21 col2\" >3</td>\n",
       "      <td id=\"T_a7ce0_row21_col3\" class=\"data row21 col3\" >0.974390</td>\n",
       "      <td id=\"T_a7ce0_row21_col4\" class=\"data row21 col4\" >0.025610</td>\n",
       "      <td id=\"T_a7ce0_row21_col5\" class=\"data row21 col5\" >4.624120</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_a7ce0_level0_row22\" class=\"row_heading level0 row22\" >15</th>\n",
       "      <td id=\"T_a7ce0_row22_col0\" class=\"data row22 col0\" >3</td>\n",
       "      <td id=\"T_a7ce0_row22_col1\" class=\"data row22 col1\" >2</td>\n",
       "      <td id=\"T_a7ce0_row22_col2\" class=\"data row22 col2\" >3</td>\n",
       "      <td id=\"T_a7ce0_row22_col3\" class=\"data row22 col3\" >0.974390</td>\n",
       "      <td id=\"T_a7ce0_row22_col4\" class=\"data row22 col4\" >0.025610</td>\n",
       "      <td id=\"T_a7ce0_row22_col5\" class=\"data row22 col5\" >5.750613</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_a7ce0_level0_row23\" class=\"row_heading level0 row23\" >26</th>\n",
       "      <td id=\"T_a7ce0_row23_col0\" class=\"data row23 col0\" >3</td>\n",
       "      <td id=\"T_a7ce0_row23_col1\" class=\"data row23 col1\" >3</td>\n",
       "      <td id=\"T_a7ce0_row23_col2\" class=\"data row23 col2\" >2</td>\n",
       "      <td id=\"T_a7ce0_row23_col3\" class=\"data row23 col3\" >0.974390</td>\n",
       "      <td id=\"T_a7ce0_row23_col4\" class=\"data row23 col4\" >0.025610</td>\n",
       "      <td id=\"T_a7ce0_row23_col5\" class=\"data row23 col5\" >5.931033</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_a7ce0_level0_row24\" class=\"row_heading level0 row24\" >28</th>\n",
       "      <td id=\"T_a7ce0_row24_col0\" class=\"data row24 col0\" >4</td>\n",
       "      <td id=\"T_a7ce0_row24_col1\" class=\"data row24 col1\" >3</td>\n",
       "      <td id=\"T_a7ce0_row24_col2\" class=\"data row24 col2\" >2</td>\n",
       "      <td id=\"T_a7ce0_row24_col3\" class=\"data row24 col3\" >0.974390</td>\n",
       "      <td id=\"T_a7ce0_row24_col4\" class=\"data row24 col4\" >0.025610</td>\n",
       "      <td id=\"T_a7ce0_row24_col5\" class=\"data row24 col5\" >2.392375</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_a7ce0_level0_row25\" class=\"row_heading level0 row25\" >29</th>\n",
       "      <td id=\"T_a7ce0_row25_col0\" class=\"data row25 col0\" >4</td>\n",
       "      <td id=\"T_a7ce0_row25_col1\" class=\"data row25 col1\" >3</td>\n",
       "      <td id=\"T_a7ce0_row25_col2\" class=\"data row25 col2\" >3</td>\n",
       "      <td id=\"T_a7ce0_row25_col3\" class=\"data row25 col3\" >0.974390</td>\n",
       "      <td id=\"T_a7ce0_row25_col4\" class=\"data row25 col4\" >0.025610</td>\n",
       "      <td id=\"T_a7ce0_row25_col5\" class=\"data row25 col5\" >3.872352</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_a7ce0_level0_row26\" class=\"row_heading level0 row26\" >2</th>\n",
       "      <td id=\"T_a7ce0_row26_col0\" class=\"data row26 col0\" >3</td>\n",
       "      <td id=\"T_a7ce0_row26_col1\" class=\"data row26 col1\" >1</td>\n",
       "      <td id=\"T_a7ce0_row26_col2\" class=\"data row26 col2\" >2</td>\n",
       "      <td id=\"T_a7ce0_row26_col3\" class=\"data row26 col3\" >0.973171</td>\n",
       "      <td id=\"T_a7ce0_row26_col4\" class=\"data row26 col4\" >0.026829</td>\n",
       "      <td id=\"T_a7ce0_row26_col5\" class=\"data row26 col5\" >0.935311</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_a7ce0_level0_row27\" class=\"row_heading level0 row27\" >16</th>\n",
       "      <td id=\"T_a7ce0_row27_col0\" class=\"data row27 col0\" >4</td>\n",
       "      <td id=\"T_a7ce0_row27_col1\" class=\"data row27 col1\" >2</td>\n",
       "      <td id=\"T_a7ce0_row27_col2\" class=\"data row27 col2\" >2</td>\n",
       "      <td id=\"T_a7ce0_row27_col3\" class=\"data row27 col3\" >0.973171</td>\n",
       "      <td id=\"T_a7ce0_row27_col4\" class=\"data row27 col4\" >0.026829</td>\n",
       "      <td id=\"T_a7ce0_row27_col5\" class=\"data row27 col5\" >1.954091</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_a7ce0_level0_row28\" class=\"row_heading level0 row28\" >27</th>\n",
       "      <td id=\"T_a7ce0_row28_col0\" class=\"data row28 col0\" >3</td>\n",
       "      <td id=\"T_a7ce0_row28_col1\" class=\"data row28 col1\" >3</td>\n",
       "      <td id=\"T_a7ce0_row28_col2\" class=\"data row28 col2\" >3</td>\n",
       "      <td id=\"T_a7ce0_row28_col3\" class=\"data row28 col3\" >0.973171</td>\n",
       "      <td id=\"T_a7ce0_row28_col4\" class=\"data row28 col4\" >0.026829</td>\n",
       "      <td id=\"T_a7ce0_row28_col5\" class=\"data row28 col5\" >4.064955</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_a7ce0_level0_row29\" class=\"row_heading level0 row29\" >14</th>\n",
       "      <td id=\"T_a7ce0_row29_col0\" class=\"data row29 col0\" >3</td>\n",
       "      <td id=\"T_a7ce0_row29_col1\" class=\"data row29 col1\" >2</td>\n",
       "      <td id=\"T_a7ce0_row29_col2\" class=\"data row29 col2\" >2</td>\n",
       "      <td id=\"T_a7ce0_row29_col3\" class=\"data row29 col3\" >0.973171</td>\n",
       "      <td id=\"T_a7ce0_row29_col4\" class=\"data row29 col4\" >0.026829</td>\n",
       "      <td id=\"T_a7ce0_row29_col5\" class=\"data row29 col5\" >1.891534</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_a7ce0_level0_row30\" class=\"row_heading level0 row30\" >25</th>\n",
       "      <td id=\"T_a7ce0_row30_col0\" class=\"data row30 col0\" >2</td>\n",
       "      <td id=\"T_a7ce0_row30_col1\" class=\"data row30 col1\" >3</td>\n",
       "      <td id=\"T_a7ce0_row30_col2\" class=\"data row30 col2\" >3</td>\n",
       "      <td id=\"T_a7ce0_row30_col3\" class=\"data row30 col3\" >0.950000</td>\n",
       "      <td id=\"T_a7ce0_row30_col4\" class=\"data row30 col4\" >0.050000</td>\n",
       "      <td id=\"T_a7ce0_row30_col5\" class=\"data row30 col5\" >11.395488</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_a7ce0_level0_row31\" class=\"row_heading level0 row31\" >24</th>\n",
       "      <td id=\"T_a7ce0_row31_col0\" class=\"data row31 col0\" >2</td>\n",
       "      <td id=\"T_a7ce0_row31_col1\" class=\"data row31 col1\" >3</td>\n",
       "      <td id=\"T_a7ce0_row31_col2\" class=\"data row31 col2\" >2</td>\n",
       "      <td id=\"T_a7ce0_row31_col3\" class=\"data row31 col3\" >0.950000</td>\n",
       "      <td id=\"T_a7ce0_row31_col4\" class=\"data row31 col4\" >0.050000</td>\n",
       "      <td id=\"T_a7ce0_row31_col5\" class=\"data row31 col5\" >2.339382</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_a7ce0_level0_row32\" class=\"row_heading level0 row32\" >12</th>\n",
       "      <td id=\"T_a7ce0_row32_col0\" class=\"data row32 col0\" >2</td>\n",
       "      <td id=\"T_a7ce0_row32_col1\" class=\"data row32 col1\" >2</td>\n",
       "      <td id=\"T_a7ce0_row32_col2\" class=\"data row32 col2\" >2</td>\n",
       "      <td id=\"T_a7ce0_row32_col3\" class=\"data row32 col3\" >0.928049</td>\n",
       "      <td id=\"T_a7ce0_row32_col4\" class=\"data row32 col4\" >0.047561</td>\n",
       "      <td id=\"T_a7ce0_row32_col5\" class=\"data row32 col5\" >1.803892</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_a7ce0_level0_row33\" class=\"row_heading level0 row33\" >13</th>\n",
       "      <td id=\"T_a7ce0_row33_col0\" class=\"data row33 col0\" >2</td>\n",
       "      <td id=\"T_a7ce0_row33_col1\" class=\"data row33 col1\" >2</td>\n",
       "      <td id=\"T_a7ce0_row33_col2\" class=\"data row33 col2\" >3</td>\n",
       "      <td id=\"T_a7ce0_row33_col3\" class=\"data row33 col3\" >0.928049</td>\n",
       "      <td id=\"T_a7ce0_row33_col4\" class=\"data row33 col4\" >0.047561</td>\n",
       "      <td id=\"T_a7ce0_row33_col5\" class=\"data row33 col5\" >2.925088</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_a7ce0_level0_row34\" class=\"row_heading level0 row34\" >1</th>\n",
       "      <td id=\"T_a7ce0_row34_col0\" class=\"data row34 col0\" >2</td>\n",
       "      <td id=\"T_a7ce0_row34_col1\" class=\"data row34 col1\" >1</td>\n",
       "      <td id=\"T_a7ce0_row34_col2\" class=\"data row34 col2\" >3</td>\n",
       "      <td id=\"T_a7ce0_row34_col3\" class=\"data row34 col3\" >0.793902</td>\n",
       "      <td id=\"T_a7ce0_row34_col4\" class=\"data row34 col4\" >0.035366</td>\n",
       "      <td id=\"T_a7ce0_row34_col5\" class=\"data row34 col5\" >1.626198</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_a7ce0_level0_row35\" class=\"row_heading level0 row35\" >0</th>\n",
       "      <td id=\"T_a7ce0_row35_col0\" class=\"data row35 col0\" >2</td>\n",
       "      <td id=\"T_a7ce0_row35_col1\" class=\"data row35 col1\" >1</td>\n",
       "      <td id=\"T_a7ce0_row35_col2\" class=\"data row35 col2\" >2</td>\n",
       "      <td id=\"T_a7ce0_row35_col3\" class=\"data row35 col3\" >0.793902</td>\n",
       "      <td id=\"T_a7ce0_row35_col4\" class=\"data row35 col4\" >0.035366</td>\n",
       "      <td id=\"T_a7ce0_row35_col5\" class=\"data row35 col5\" >1.136223</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n"
      ],
      "text/plain": [
       "<pandas.io.formats.style.Styler at 0x7fa56690d4b0>"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# XGB\n",
    "\n",
    "index = 2\n",
    "xgb_results.style.background_gradient(\n",
    "    subset=[\"n_bits\", \"mean_test_score\", \"fhe_inference_time\"]\n",
    ").apply(\n",
    "    lambda index: [\"background-color: green\"] * len(index), axis=1, subset=pd.IndexSlice[[index], :]\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<style type=\"text/css\">\n",
       "#T_eeca8_row0_col1, #T_eeca8_row1_col1, #T_eeca8_row2_col1, #T_eeca8_row15_col1, #T_eeca8_row15_col6, #T_eeca8_row16_col1, #T_eeca8_row17_col1, #T_eeca8_row18_col1, #T_eeca8_row30_col1, #T_eeca8_row32_col1, #T_eeca8_row35_col1, #T_eeca8_row36_col1, #T_eeca8_row37_col1, #T_eeca8_row48_col4, #T_eeca8_row49_col1, #T_eeca8_row49_col4, #T_eeca8_row50_col1, #T_eeca8_row50_col4, #T_eeca8_row51_col4, #T_eeca8_row52_col4, #T_eeca8_row53_col4, #T_eeca8_row54_col4, #T_eeca8_row55_col1, #T_eeca8_row55_col4, #T_eeca8_row56_col4, #T_eeca8_row57_col4, #T_eeca8_row58_col4, #T_eeca8_row59_col4, #T_eeca8_row60_col4, #T_eeca8_row61_col4, #T_eeca8_row62_col4, #T_eeca8_row63_col1, #T_eeca8_row63_col4 {\n",
       "  background-color: #fff7fb;\n",
       "  color: #000000;\n",
       "}\n",
       "#T_eeca8_row0_col4, #T_eeca8_row1_col4, #T_eeca8_row2_col4, #T_eeca8_row3_col4, #T_eeca8_row4_col4, #T_eeca8_row5_col4, #T_eeca8_row6_col4, #T_eeca8_row7_col4, #T_eeca8_row8_col4, #T_eeca8_row10_col4, #T_eeca8_row11_col1, #T_eeca8_row11_col4, #T_eeca8_row12_col1, #T_eeca8_row12_col4, #T_eeca8_row13_col1, #T_eeca8_row13_col4, #T_eeca8_row14_col1, #T_eeca8_row14_col4, #T_eeca8_row15_col4, #T_eeca8_row16_col4, #T_eeca8_row17_col4, #T_eeca8_row18_col4, #T_eeca8_row19_col4, #T_eeca8_row20_col4, #T_eeca8_row21_col4, #T_eeca8_row22_col4, #T_eeca8_row23_col4, #T_eeca8_row24_col4, #T_eeca8_row25_col4, #T_eeca8_row26_col4, #T_eeca8_row27_col1, #T_eeca8_row27_col4, #T_eeca8_row28_col1, #T_eeca8_row28_col4, #T_eeca8_row29_col1, #T_eeca8_row29_col4, #T_eeca8_row30_col4, #T_eeca8_row31_col1, #T_eeca8_row31_col4, #T_eeca8_row32_col4, #T_eeca8_row33_col4, #T_eeca8_row34_col1, #T_eeca8_row34_col4, #T_eeca8_row35_col4, #T_eeca8_row36_col4, #T_eeca8_row37_col4, #T_eeca8_row38_col4, #T_eeca8_row39_col4, #T_eeca8_row40_col4, #T_eeca8_row41_col4, #T_eeca8_row42_col4, #T_eeca8_row43_col4, #T_eeca8_row44_col4, #T_eeca8_row45_col1, #T_eeca8_row45_col4, #T_eeca8_row46_col1, #T_eeca8_row46_col4, #T_eeca8_row47_col1, #T_eeca8_row47_col4, #T_eeca8_row48_col6, #T_eeca8_row59_col1, #T_eeca8_row60_col1, #T_eeca8_row61_col1, #T_eeca8_row62_col1 {\n",
       "  background-color: #023858;\n",
       "  color: #f1f1f1;\n",
       "}\n",
       "#T_eeca8_row0_col6 {\n",
       "  background-color: #f9f2f8;\n",
       "  color: #000000;\n",
       "}\n",
       "#T_eeca8_row1_col6, #T_eeca8_row49_col6 {\n",
       "  background-color: #f6eff7;\n",
       "  color: #000000;\n",
       "}\n",
       "#T_eeca8_row2_col6, #T_eeca8_row39_col6, #T_eeca8_row41_col6 {\n",
       "  background-color: #ede7f2;\n",
       "  color: #000000;\n",
       "}\n",
       "#T_eeca8_row3_col1, #T_eeca8_row4_col1, #T_eeca8_row5_col1, #T_eeca8_row6_col1, #T_eeca8_row19_col1, #T_eeca8_row20_col1, #T_eeca8_row21_col1, #T_eeca8_row22_col1, #T_eeca8_row33_col1, #T_eeca8_row38_col1, #T_eeca8_row39_col1, #T_eeca8_row40_col1, #T_eeca8_row51_col1, #T_eeca8_row52_col1, #T_eeca8_row53_col1, #T_eeca8_row54_col1 {\n",
       "  background-color: #91b5d6;\n",
       "  color: #000000;\n",
       "}\n",
       "#T_eeca8_row3_col6 {\n",
       "  background-color: #c1cae2;\n",
       "  color: #000000;\n",
       "}\n",
       "#T_eeca8_row4_col6 {\n",
       "  background-color: #e4e1ef;\n",
       "  color: #000000;\n",
       "}\n",
       "#T_eeca8_row5_col6, #T_eeca8_row44_col6 {\n",
       "  background-color: #e1dfed;\n",
       "  color: #000000;\n",
       "}\n",
       "#T_eeca8_row6_col6, #T_eeca8_row40_col6 {\n",
       "  background-color: #e6e2ef;\n",
       "  color: #000000;\n",
       "}\n",
       "#T_eeca8_row7_col1, #T_eeca8_row8_col1, #T_eeca8_row10_col1, #T_eeca8_row23_col1, #T_eeca8_row24_col1, #T_eeca8_row25_col1, #T_eeca8_row26_col1, #T_eeca8_row41_col1, #T_eeca8_row42_col1, #T_eeca8_row43_col1, #T_eeca8_row44_col1, #T_eeca8_row48_col1, #T_eeca8_row56_col1, #T_eeca8_row57_col1, #T_eeca8_row58_col1 {\n",
       "  background-color: #045d92;\n",
       "  color: #f1f1f1;\n",
       "}\n",
       "#T_eeca8_row7_col6, #T_eeca8_row8_col6, #T_eeca8_row56_col6 {\n",
       "  background-color: #e9e5f1;\n",
       "  color: #000000;\n",
       "}\n",
       "#T_eeca8_row9_col0, #T_eeca8_row9_col2, #T_eeca8_row9_col3, #T_eeca8_row9_col5 {\n",
       "  background-color: green;\n",
       "}\n",
       "#T_eeca8_row9_col1 {\n",
       "  background-color: #045d92;\n",
       "  color: #f1f1f1;\n",
       "  background-color: green;\n",
       "}\n",
       "#T_eeca8_row9_col4 {\n",
       "  background-color: #023858;\n",
       "  color: #f1f1f1;\n",
       "  background-color: green;\n",
       "}\n",
       "#T_eeca8_row9_col6 {\n",
       "  background-color: #ece7f2;\n",
       "  color: #000000;\n",
       "  background-color: green;\n",
       "}\n",
       "#T_eeca8_row10_col6, #T_eeca8_row21_col6 {\n",
       "  background-color: #ede8f3;\n",
       "  color: #000000;\n",
       "}\n",
       "#T_eeca8_row11_col6, #T_eeca8_row13_col6, #T_eeca8_row31_col6, #T_eeca8_row59_col6 {\n",
       "  background-color: #d9d8ea;\n",
       "  color: #000000;\n",
       "}\n",
       "#T_eeca8_row12_col6 {\n",
       "  background-color: #d0d1e6;\n",
       "  color: #000000;\n",
       "}\n",
       "#T_eeca8_row14_col6, #T_eeca8_row47_col6 {\n",
       "  background-color: #dcdaeb;\n",
       "  color: #000000;\n",
       "}\n",
       "#T_eeca8_row16_col6, #T_eeca8_row30_col6 {\n",
       "  background-color: #faf3f9;\n",
       "  color: #000000;\n",
       "}\n",
       "#T_eeca8_row17_col6 {\n",
       "  background-color: #fcf4fa;\n",
       "  color: #000000;\n",
       "}\n",
       "#T_eeca8_row18_col6 {\n",
       "  background-color: #fbf3f9;\n",
       "  color: #000000;\n",
       "}\n",
       "#T_eeca8_row19_col6, #T_eeca8_row25_col6, #T_eeca8_row38_col6 {\n",
       "  background-color: #eee8f3;\n",
       "  color: #000000;\n",
       "}\n",
       "#T_eeca8_row20_col6 {\n",
       "  background-color: #d6d6e9;\n",
       "  color: #000000;\n",
       "}\n",
       "#T_eeca8_row22_col6, #T_eeca8_row24_col6, #T_eeca8_row57_col6 {\n",
       "  background-color: #ebe6f2;\n",
       "  color: #000000;\n",
       "}\n",
       "#T_eeca8_row23_col6 {\n",
       "  background-color: #e3e0ee;\n",
       "  color: #000000;\n",
       "}\n",
       "#T_eeca8_row26_col6, #T_eeca8_row33_col6 {\n",
       "  background-color: #ece7f2;\n",
       "  color: #000000;\n",
       "}\n",
       "#T_eeca8_row27_col6 {\n",
       "  background-color: #dbdaeb;\n",
       "  color: #000000;\n",
       "}\n",
       "#T_eeca8_row28_col6 {\n",
       "  background-color: #c4cbe3;\n",
       "  color: #000000;\n",
       "}\n",
       "#T_eeca8_row29_col6 {\n",
       "  background-color: #d5d5e8;\n",
       "  color: #000000;\n",
       "}\n",
       "#T_eeca8_row32_col6 {\n",
       "  background-color: #c5cce3;\n",
       "  color: #000000;\n",
       "}\n",
       "#T_eeca8_row34_col6 {\n",
       "  background-color: #b5c4df;\n",
       "  color: #000000;\n",
       "}\n",
       "#T_eeca8_row35_col6 {\n",
       "  background-color: #f8f1f8;\n",
       "  color: #000000;\n",
       "}\n",
       "#T_eeca8_row36_col6 {\n",
       "  background-color: #f4eef6;\n",
       "  color: #000000;\n",
       "}\n",
       "#T_eeca8_row37_col6 {\n",
       "  background-color: #f1ebf4;\n",
       "  color: #000000;\n",
       "}\n",
       "#T_eeca8_row42_col6 {\n",
       "  background-color: #adc1dd;\n",
       "  color: #000000;\n",
       "}\n",
       "#T_eeca8_row43_col6 {\n",
       "  background-color: #eae6f1;\n",
       "  color: #000000;\n",
       "}\n",
       "#T_eeca8_row45_col6 {\n",
       "  background-color: #96b6d7;\n",
       "  color: #000000;\n",
       "}\n",
       "#T_eeca8_row46_col6 {\n",
       "  background-color: #ced0e6;\n",
       "  color: #000000;\n",
       "}\n",
       "#T_eeca8_row50_col6 {\n",
       "  background-color: #f4edf6;\n",
       "  color: #000000;\n",
       "}\n",
       "#T_eeca8_row51_col6 {\n",
       "  background-color: #eee9f3;\n",
       "  color: #000000;\n",
       "}\n",
       "#T_eeca8_row52_col6 {\n",
       "  background-color: #f0eaf4;\n",
       "  color: #000000;\n",
       "}\n",
       "#T_eeca8_row53_col6 {\n",
       "  background-color: #e7e3f0;\n",
       "  color: #000000;\n",
       "}\n",
       "#T_eeca8_row54_col6 {\n",
       "  background-color: #d3d4e7;\n",
       "  color: #000000;\n",
       "}\n",
       "#T_eeca8_row55_col6 {\n",
       "  background-color: #f5eff6;\n",
       "  color: #000000;\n",
       "}\n",
       "#T_eeca8_row58_col6 {\n",
       "  background-color: #e8e4f0;\n",
       "  color: #000000;\n",
       "}\n",
       "#T_eeca8_row60_col6 {\n",
       "  background-color: #dfddec;\n",
       "  color: #000000;\n",
       "}\n",
       "#T_eeca8_row61_col6 {\n",
       "  background-color: #c6cce3;\n",
       "  color: #000000;\n",
       "}\n",
       "#T_eeca8_row62_col6 {\n",
       "  background-color: #bbc7e0;\n",
       "  color: #000000;\n",
       "}\n",
       "#T_eeca8_row63_col6 {\n",
       "  background-color: #f1ebf5;\n",
       "  color: #000000;\n",
       "}\n",
       "</style>\n",
       "<table id=\"T_eeca8\">\n",
       "  <thead>\n",
       "    <tr>\n",
       "      <th class=\"blank level0\" >&nbsp;</th>\n",
       "      <th id=\"T_eeca8_level0_col0\" class=\"col_heading level0 col0\" >C</th>\n",
       "      <th id=\"T_eeca8_level0_col1\" class=\"col_heading level0 col1\" >n_bits</th>\n",
       "      <th id=\"T_eeca8_level0_col2\" class=\"col_heading level0 col2\" >solver</th>\n",
       "      <th id=\"T_eeca8_level0_col3\" class=\"col_heading level0 col3\" >multi_class</th>\n",
       "      <th id=\"T_eeca8_level0_col4\" class=\"col_heading level0 col4\" >mean_test_score</th>\n",
       "      <th id=\"T_eeca8_level0_col5\" class=\"col_heading level0 col5\" >std_test_score</th>\n",
       "      <th id=\"T_eeca8_level0_col6\" class=\"col_heading level0 col6\" >fhe_inference_time</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th id=\"T_eeca8_level0_row0\" class=\"row_heading level0 row0\" >32</th>\n",
       "      <td id=\"T_eeca8_row0_col0\" class=\"data row0 col0\" >0.900000</td>\n",
       "      <td id=\"T_eeca8_row0_col1\" class=\"data row0 col1\" >7</td>\n",
       "      <td id=\"T_eeca8_row0_col2\" class=\"data row0 col2\" >newton-cg</td>\n",
       "      <td id=\"T_eeca8_row0_col3\" class=\"data row0 col3\" >auto</td>\n",
       "      <td id=\"T_eeca8_row0_col4\" class=\"data row0 col4\" >1.000000</td>\n",
       "      <td id=\"T_eeca8_row0_col5\" class=\"data row0 col5\" >0.000000</td>\n",
       "      <td id=\"T_eeca8_row0_col6\" class=\"data row0 col6\" >0.005855</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_eeca8_level0_row1\" class=\"row_heading level0 row1\" >48</th>\n",
       "      <td id=\"T_eeca8_row1_col0\" class=\"data row1 col0\" >1.000000</td>\n",
       "      <td id=\"T_eeca8_row1_col1\" class=\"data row1 col1\" >7</td>\n",
       "      <td id=\"T_eeca8_row1_col2\" class=\"data row1 col2\" >newton-cg</td>\n",
       "      <td id=\"T_eeca8_row1_col3\" class=\"data row1 col3\" >auto</td>\n",
       "      <td id=\"T_eeca8_row1_col4\" class=\"data row1 col4\" >1.000000</td>\n",
       "      <td id=\"T_eeca8_row1_col5\" class=\"data row1 col5\" >0.000000</td>\n",
       "      <td id=\"T_eeca8_row1_col6\" class=\"data row1 col6\" >0.006321</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_eeca8_level0_row2\" class=\"row_heading level0 row2\" >35</th>\n",
       "      <td id=\"T_eeca8_row2_col0\" class=\"data row2 col0\" >0.900000</td>\n",
       "      <td id=\"T_eeca8_row2_col1\" class=\"data row2 col1\" >7</td>\n",
       "      <td id=\"T_eeca8_row2_col2\" class=\"data row2 col2\" >lbfgs</td>\n",
       "      <td id=\"T_eeca8_row2_col3\" class=\"data row2 col3\" >auto</td>\n",
       "      <td id=\"T_eeca8_row2_col4\" class=\"data row2 col4\" >1.000000</td>\n",
       "      <td id=\"T_eeca8_row2_col5\" class=\"data row2 col5\" >0.000000</td>\n",
       "      <td id=\"T_eeca8_row2_col6\" class=\"data row2 col6\" >0.007907</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_eeca8_level0_row3\" class=\"row_heading level0 row3\" >36</th>\n",
       "      <td id=\"T_eeca8_row3_col0\" class=\"data row3 col0\" >0.900000</td>\n",
       "      <td id=\"T_eeca8_row3_col1\" class=\"data row3 col1\" >10</td>\n",
       "      <td id=\"T_eeca8_row3_col2\" class=\"data row3 col2\" >newton-cg</td>\n",
       "      <td id=\"T_eeca8_row3_col3\" class=\"data row3 col3\" >auto</td>\n",
       "      <td id=\"T_eeca8_row3_col4\" class=\"data row3 col4\" >1.000000</td>\n",
       "      <td id=\"T_eeca8_row3_col5\" class=\"data row3 col5\" >0.000000</td>\n",
       "      <td id=\"T_eeca8_row3_col6\" class=\"data row3 col6\" >0.012095</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_eeca8_level0_row4\" class=\"row_heading level0 row4\" >37</th>\n",
       "      <td id=\"T_eeca8_row4_col0\" class=\"data row4 col0\" >0.900000</td>\n",
       "      <td id=\"T_eeca8_row4_col1\" class=\"data row4 col1\" >10</td>\n",
       "      <td id=\"T_eeca8_row4_col2\" class=\"data row4 col2\" >sag</td>\n",
       "      <td id=\"T_eeca8_row4_col3\" class=\"data row4 col3\" >auto</td>\n",
       "      <td id=\"T_eeca8_row4_col4\" class=\"data row4 col4\" >1.000000</td>\n",
       "      <td id=\"T_eeca8_row4_col5\" class=\"data row4 col5\" >0.000000</td>\n",
       "      <td id=\"T_eeca8_row4_col6\" class=\"data row4 col6\" >0.008803</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_eeca8_level0_row5\" class=\"row_heading level0 row5\" >38</th>\n",
       "      <td id=\"T_eeca8_row5_col0\" class=\"data row5 col0\" >0.900000</td>\n",
       "      <td id=\"T_eeca8_row5_col1\" class=\"data row5 col1\" >10</td>\n",
       "      <td id=\"T_eeca8_row5_col2\" class=\"data row5 col2\" >saga</td>\n",
       "      <td id=\"T_eeca8_row5_col3\" class=\"data row5 col3\" >auto</td>\n",
       "      <td id=\"T_eeca8_row5_col4\" class=\"data row5 col4\" >1.000000</td>\n",
       "      <td id=\"T_eeca8_row5_col5\" class=\"data row5 col5\" >0.000000</td>\n",
       "      <td id=\"T_eeca8_row5_col6\" class=\"data row5 col6\" >0.009139</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_eeca8_level0_row6\" class=\"row_heading level0 row6\" >39</th>\n",
       "      <td id=\"T_eeca8_row6_col0\" class=\"data row6 col0\" >0.900000</td>\n",
       "      <td id=\"T_eeca8_row6_col1\" class=\"data row6 col1\" >10</td>\n",
       "      <td id=\"T_eeca8_row6_col2\" class=\"data row6 col2\" >lbfgs</td>\n",
       "      <td id=\"T_eeca8_row6_col3\" class=\"data row6 col3\" >auto</td>\n",
       "      <td id=\"T_eeca8_row6_col4\" class=\"data row6 col4\" >1.000000</td>\n",
       "      <td id=\"T_eeca8_row6_col5\" class=\"data row6 col5\" >0.000000</td>\n",
       "      <td id=\"T_eeca8_row6_col6\" class=\"data row6 col6\" >0.008595</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_eeca8_level0_row7\" class=\"row_heading level0 row7\" >40</th>\n",
       "      <td id=\"T_eeca8_row7_col0\" class=\"data row7 col0\" >0.900000</td>\n",
       "      <td id=\"T_eeca8_row7_col1\" class=\"data row7 col1\" >13</td>\n",
       "      <td id=\"T_eeca8_row7_col2\" class=\"data row7 col2\" >newton-cg</td>\n",
       "      <td id=\"T_eeca8_row7_col3\" class=\"data row7 col3\" >auto</td>\n",
       "      <td id=\"T_eeca8_row7_col4\" class=\"data row7 col4\" >1.000000</td>\n",
       "      <td id=\"T_eeca8_row7_col5\" class=\"data row7 col5\" >0.000000</td>\n",
       "      <td id=\"T_eeca8_row7_col6\" class=\"data row7 col6\" >0.008255</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_eeca8_level0_row8\" class=\"row_heading level0 row8\" >41</th>\n",
       "      <td id=\"T_eeca8_row8_col0\" class=\"data row8 col0\" >0.900000</td>\n",
       "      <td id=\"T_eeca8_row8_col1\" class=\"data row8 col1\" >13</td>\n",
       "      <td id=\"T_eeca8_row8_col2\" class=\"data row8 col2\" >sag</td>\n",
       "      <td id=\"T_eeca8_row8_col3\" class=\"data row8 col3\" >auto</td>\n",
       "      <td id=\"T_eeca8_row8_col4\" class=\"data row8 col4\" >1.000000</td>\n",
       "      <td id=\"T_eeca8_row8_col5\" class=\"data row8 col5\" >0.000000</td>\n",
       "      <td id=\"T_eeca8_row8_col6\" class=\"data row8 col6\" >0.008248</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_eeca8_level0_row9\" class=\"row_heading level0 row9\" >42</th>\n",
       "      <td id=\"T_eeca8_row9_col0\" class=\"data row9 col0\" >0.900000</td>\n",
       "      <td id=\"T_eeca8_row9_col1\" class=\"data row9 col1\" >13</td>\n",
       "      <td id=\"T_eeca8_row9_col2\" class=\"data row9 col2\" >saga</td>\n",
       "      <td id=\"T_eeca8_row9_col3\" class=\"data row9 col3\" >auto</td>\n",
       "      <td id=\"T_eeca8_row9_col4\" class=\"data row9 col4\" >1.000000</td>\n",
       "      <td id=\"T_eeca8_row9_col5\" class=\"data row9 col5\" >0.000000</td>\n",
       "      <td id=\"T_eeca8_row9_col6\" class=\"data row9 col6\" >0.007975</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_eeca8_level0_row10\" class=\"row_heading level0 row10\" >43</th>\n",
       "      <td id=\"T_eeca8_row10_col0\" class=\"data row10 col0\" >0.900000</td>\n",
       "      <td id=\"T_eeca8_row10_col1\" class=\"data row10 col1\" >13</td>\n",
       "      <td id=\"T_eeca8_row10_col2\" class=\"data row10 col2\" >lbfgs</td>\n",
       "      <td id=\"T_eeca8_row10_col3\" class=\"data row10 col3\" >auto</td>\n",
       "      <td id=\"T_eeca8_row10_col4\" class=\"data row10 col4\" >1.000000</td>\n",
       "      <td id=\"T_eeca8_row10_col5\" class=\"data row10 col5\" >0.000000</td>\n",
       "      <td id=\"T_eeca8_row10_col6\" class=\"data row10 col6\" >0.007803</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_eeca8_level0_row11\" class=\"row_heading level0 row11\" >44</th>\n",
       "      <td id=\"T_eeca8_row11_col0\" class=\"data row11 col0\" >0.900000</td>\n",
       "      <td id=\"T_eeca8_row11_col1\" class=\"data row11 col1\" >14</td>\n",
       "      <td id=\"T_eeca8_row11_col2\" class=\"data row11 col2\" >newton-cg</td>\n",
       "      <td id=\"T_eeca8_row11_col3\" class=\"data row11 col3\" >auto</td>\n",
       "      <td id=\"T_eeca8_row11_col4\" class=\"data row11 col4\" >1.000000</td>\n",
       "      <td id=\"T_eeca8_row11_col5\" class=\"data row11 col5\" >0.000000</td>\n",
       "      <td id=\"T_eeca8_row11_col6\" class=\"data row11 col6\" >0.010030</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_eeca8_level0_row12\" class=\"row_heading level0 row12\" >45</th>\n",
       "      <td id=\"T_eeca8_row12_col0\" class=\"data row12 col0\" >0.900000</td>\n",
       "      <td id=\"T_eeca8_row12_col1\" class=\"data row12 col1\" >14</td>\n",
       "      <td id=\"T_eeca8_row12_col2\" class=\"data row12 col2\" >sag</td>\n",
       "      <td id=\"T_eeca8_row12_col3\" class=\"data row12 col3\" >auto</td>\n",
       "      <td id=\"T_eeca8_row12_col4\" class=\"data row12 col4\" >1.000000</td>\n",
       "      <td id=\"T_eeca8_row12_col5\" class=\"data row12 col5\" >0.000000</td>\n",
       "      <td id=\"T_eeca8_row12_col6\" class=\"data row12 col6\" >0.011017</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_eeca8_level0_row13\" class=\"row_heading level0 row13\" >46</th>\n",
       "      <td id=\"T_eeca8_row13_col0\" class=\"data row13 col0\" >0.900000</td>\n",
       "      <td id=\"T_eeca8_row13_col1\" class=\"data row13 col1\" >14</td>\n",
       "      <td id=\"T_eeca8_row13_col2\" class=\"data row13 col2\" >saga</td>\n",
       "      <td id=\"T_eeca8_row13_col3\" class=\"data row13 col3\" >auto</td>\n",
       "      <td id=\"T_eeca8_row13_col4\" class=\"data row13 col4\" >1.000000</td>\n",
       "      <td id=\"T_eeca8_row13_col5\" class=\"data row13 col5\" >0.000000</td>\n",
       "      <td id=\"T_eeca8_row13_col6\" class=\"data row13 col6\" >0.009961</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_eeca8_level0_row14\" class=\"row_heading level0 row14\" >47</th>\n",
       "      <td id=\"T_eeca8_row14_col0\" class=\"data row14 col0\" >0.900000</td>\n",
       "      <td id=\"T_eeca8_row14_col1\" class=\"data row14 col1\" >14</td>\n",
       "      <td id=\"T_eeca8_row14_col2\" class=\"data row14 col2\" >lbfgs</td>\n",
       "      <td id=\"T_eeca8_row14_col3\" class=\"data row14 col3\" >auto</td>\n",
       "      <td id=\"T_eeca8_row14_col4\" class=\"data row14 col4\" >1.000000</td>\n",
       "      <td id=\"T_eeca8_row14_col5\" class=\"data row14 col5\" >0.000000</td>\n",
       "      <td id=\"T_eeca8_row14_col6\" class=\"data row14 col6\" >0.009642</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_eeca8_level0_row15\" class=\"row_heading level0 row15\" >49</th>\n",
       "      <td id=\"T_eeca8_row15_col0\" class=\"data row15 col0\" >1.000000</td>\n",
       "      <td id=\"T_eeca8_row15_col1\" class=\"data row15 col1\" >7</td>\n",
       "      <td id=\"T_eeca8_row15_col2\" class=\"data row15 col2\" >sag</td>\n",
       "      <td id=\"T_eeca8_row15_col3\" class=\"data row15 col3\" >auto</td>\n",
       "      <td id=\"T_eeca8_row15_col4\" class=\"data row15 col4\" >1.000000</td>\n",
       "      <td id=\"T_eeca8_row15_col5\" class=\"data row15 col5\" >0.000000</td>\n",
       "      <td id=\"T_eeca8_row15_col6\" class=\"data row15 col6\" >0.004868</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_eeca8_level0_row16\" class=\"row_heading level0 row16\" >33</th>\n",
       "      <td id=\"T_eeca8_row16_col0\" class=\"data row16 col0\" >0.900000</td>\n",
       "      <td id=\"T_eeca8_row16_col1\" class=\"data row16 col1\" >7</td>\n",
       "      <td id=\"T_eeca8_row16_col2\" class=\"data row16 col2\" >sag</td>\n",
       "      <td id=\"T_eeca8_row16_col3\" class=\"data row16 col3\" >auto</td>\n",
       "      <td id=\"T_eeca8_row16_col4\" class=\"data row16 col4\" >1.000000</td>\n",
       "      <td id=\"T_eeca8_row16_col5\" class=\"data row16 col5\" >0.000000</td>\n",
       "      <td id=\"T_eeca8_row16_col6\" class=\"data row16 col6\" >0.005669</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_eeca8_level0_row17\" class=\"row_heading level0 row17\" >50</th>\n",
       "      <td id=\"T_eeca8_row17_col0\" class=\"data row17 col0\" >1.000000</td>\n",
       "      <td id=\"T_eeca8_row17_col1\" class=\"data row17 col1\" >7</td>\n",
       "      <td id=\"T_eeca8_row17_col2\" class=\"data row17 col2\" >saga</td>\n",
       "      <td id=\"T_eeca8_row17_col3\" class=\"data row17 col3\" >auto</td>\n",
       "      <td id=\"T_eeca8_row17_col4\" class=\"data row17 col4\" >1.000000</td>\n",
       "      <td id=\"T_eeca8_row17_col5\" class=\"data row17 col5\" >0.000000</td>\n",
       "      <td id=\"T_eeca8_row17_col6\" class=\"data row17 col6\" >0.005345</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_eeca8_level0_row18\" class=\"row_heading level0 row18\" >51</th>\n",
       "      <td id=\"T_eeca8_row18_col0\" class=\"data row18 col0\" >1.000000</td>\n",
       "      <td id=\"T_eeca8_row18_col1\" class=\"data row18 col1\" >7</td>\n",
       "      <td id=\"T_eeca8_row18_col2\" class=\"data row18 col2\" >lbfgs</td>\n",
       "      <td id=\"T_eeca8_row18_col3\" class=\"data row18 col3\" >auto</td>\n",
       "      <td id=\"T_eeca8_row18_col4\" class=\"data row18 col4\" >1.000000</td>\n",
       "      <td id=\"T_eeca8_row18_col5\" class=\"data row18 col5\" >0.000000</td>\n",
       "      <td id=\"T_eeca8_row18_col6\" class=\"data row18 col6\" >0.005622</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_eeca8_level0_row19\" class=\"row_heading level0 row19\" >52</th>\n",
       "      <td id=\"T_eeca8_row19_col0\" class=\"data row19 col0\" >1.000000</td>\n",
       "      <td id=\"T_eeca8_row19_col1\" class=\"data row19 col1\" >10</td>\n",
       "      <td id=\"T_eeca8_row19_col2\" class=\"data row19 col2\" >newton-cg</td>\n",
       "      <td id=\"T_eeca8_row19_col3\" class=\"data row19 col3\" >auto</td>\n",
       "      <td id=\"T_eeca8_row19_col4\" class=\"data row19 col4\" >1.000000</td>\n",
       "      <td id=\"T_eeca8_row19_col5\" class=\"data row19 col5\" >0.000000</td>\n",
       "      <td id=\"T_eeca8_row19_col6\" class=\"data row19 col6\" >0.007709</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_eeca8_level0_row20\" class=\"row_heading level0 row20\" >53</th>\n",
       "      <td id=\"T_eeca8_row20_col0\" class=\"data row20 col0\" >1.000000</td>\n",
       "      <td id=\"T_eeca8_row20_col1\" class=\"data row20 col1\" >10</td>\n",
       "      <td id=\"T_eeca8_row20_col2\" class=\"data row20 col2\" >sag</td>\n",
       "      <td id=\"T_eeca8_row20_col3\" class=\"data row20 col3\" >auto</td>\n",
       "      <td id=\"T_eeca8_row20_col4\" class=\"data row20 col4\" >1.000000</td>\n",
       "      <td id=\"T_eeca8_row20_col5\" class=\"data row20 col5\" >0.000000</td>\n",
       "      <td id=\"T_eeca8_row20_col6\" class=\"data row20 col6\" >0.010367</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_eeca8_level0_row21\" class=\"row_heading level0 row21\" >54</th>\n",
       "      <td id=\"T_eeca8_row21_col0\" class=\"data row21 col0\" >1.000000</td>\n",
       "      <td id=\"T_eeca8_row21_col1\" class=\"data row21 col1\" >10</td>\n",
       "      <td id=\"T_eeca8_row21_col2\" class=\"data row21 col2\" >saga</td>\n",
       "      <td id=\"T_eeca8_row21_col3\" class=\"data row21 col3\" >auto</td>\n",
       "      <td id=\"T_eeca8_row21_col4\" class=\"data row21 col4\" >1.000000</td>\n",
       "      <td id=\"T_eeca8_row21_col5\" class=\"data row21 col5\" >0.000000</td>\n",
       "      <td id=\"T_eeca8_row21_col6\" class=\"data row21 col6\" >0.007763</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_eeca8_level0_row22\" class=\"row_heading level0 row22\" >55</th>\n",
       "      <td id=\"T_eeca8_row22_col0\" class=\"data row22 col0\" >1.000000</td>\n",
       "      <td id=\"T_eeca8_row22_col1\" class=\"data row22 col1\" >10</td>\n",
       "      <td id=\"T_eeca8_row22_col2\" class=\"data row22 col2\" >lbfgs</td>\n",
       "      <td id=\"T_eeca8_row22_col3\" class=\"data row22 col3\" >auto</td>\n",
       "      <td id=\"T_eeca8_row22_col4\" class=\"data row22 col4\" >1.000000</td>\n",
       "      <td id=\"T_eeca8_row22_col5\" class=\"data row22 col5\" >0.000000</td>\n",
       "      <td id=\"T_eeca8_row22_col6\" class=\"data row22 col6\" >0.008013</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_eeca8_level0_row23\" class=\"row_heading level0 row23\" >56</th>\n",
       "      <td id=\"T_eeca8_row23_col0\" class=\"data row23 col0\" >1.000000</td>\n",
       "      <td id=\"T_eeca8_row23_col1\" class=\"data row23 col1\" >13</td>\n",
       "      <td id=\"T_eeca8_row23_col2\" class=\"data row23 col2\" >newton-cg</td>\n",
       "      <td id=\"T_eeca8_row23_col3\" class=\"data row23 col3\" >auto</td>\n",
       "      <td id=\"T_eeca8_row23_col4\" class=\"data row23 col4\" >1.000000</td>\n",
       "      <td id=\"T_eeca8_row23_col5\" class=\"data row23 col5\" >0.000000</td>\n",
       "      <td id=\"T_eeca8_row23_col6\" class=\"data row23 col6\" >0.008916</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_eeca8_level0_row24\" class=\"row_heading level0 row24\" >57</th>\n",
       "      <td id=\"T_eeca8_row24_col0\" class=\"data row24 col0\" >1.000000</td>\n",
       "      <td id=\"T_eeca8_row24_col1\" class=\"data row24 col1\" >13</td>\n",
       "      <td id=\"T_eeca8_row24_col2\" class=\"data row24 col2\" >sag</td>\n",
       "      <td id=\"T_eeca8_row24_col3\" class=\"data row24 col3\" >auto</td>\n",
       "      <td id=\"T_eeca8_row24_col4\" class=\"data row24 col4\" >1.000000</td>\n",
       "      <td id=\"T_eeca8_row24_col5\" class=\"data row24 col5\" >0.000000</td>\n",
       "      <td id=\"T_eeca8_row24_col6\" class=\"data row24 col6\" >0.008012</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_eeca8_level0_row25\" class=\"row_heading level0 row25\" >58</th>\n",
       "      <td id=\"T_eeca8_row25_col0\" class=\"data row25 col0\" >1.000000</td>\n",
       "      <td id=\"T_eeca8_row25_col1\" class=\"data row25 col1\" >13</td>\n",
       "      <td id=\"T_eeca8_row25_col2\" class=\"data row25 col2\" >saga</td>\n",
       "      <td id=\"T_eeca8_row25_col3\" class=\"data row25 col3\" >auto</td>\n",
       "      <td id=\"T_eeca8_row25_col4\" class=\"data row25 col4\" >1.000000</td>\n",
       "      <td id=\"T_eeca8_row25_col5\" class=\"data row25 col5\" >0.000000</td>\n",
       "      <td id=\"T_eeca8_row25_col6\" class=\"data row25 col6\" >0.007686</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_eeca8_level0_row26\" class=\"row_heading level0 row26\" >59</th>\n",
       "      <td id=\"T_eeca8_row26_col0\" class=\"data row26 col0\" >1.000000</td>\n",
       "      <td id=\"T_eeca8_row26_col1\" class=\"data row26 col1\" >13</td>\n",
       "      <td id=\"T_eeca8_row26_col2\" class=\"data row26 col2\" >lbfgs</td>\n",
       "      <td id=\"T_eeca8_row26_col3\" class=\"data row26 col3\" >auto</td>\n",
       "      <td id=\"T_eeca8_row26_col4\" class=\"data row26 col4\" >1.000000</td>\n",
       "      <td id=\"T_eeca8_row26_col5\" class=\"data row26 col5\" >0.000000</td>\n",
       "      <td id=\"T_eeca8_row26_col6\" class=\"data row26 col6\" >0.007953</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_eeca8_level0_row27\" class=\"row_heading level0 row27\" >60</th>\n",
       "      <td id=\"T_eeca8_row27_col0\" class=\"data row27 col0\" >1.000000</td>\n",
       "      <td id=\"T_eeca8_row27_col1\" class=\"data row27 col1\" >14</td>\n",
       "      <td id=\"T_eeca8_row27_col2\" class=\"data row27 col2\" >newton-cg</td>\n",
       "      <td id=\"T_eeca8_row27_col3\" class=\"data row27 col3\" >auto</td>\n",
       "      <td id=\"T_eeca8_row27_col4\" class=\"data row27 col4\" >1.000000</td>\n",
       "      <td id=\"T_eeca8_row27_col5\" class=\"data row27 col5\" >0.000000</td>\n",
       "      <td id=\"T_eeca8_row27_col6\" class=\"data row27 col6\" >0.009804</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_eeca8_level0_row28\" class=\"row_heading level0 row28\" >61</th>\n",
       "      <td id=\"T_eeca8_row28_col0\" class=\"data row28 col0\" >1.000000</td>\n",
       "      <td id=\"T_eeca8_row28_col1\" class=\"data row28 col1\" >14</td>\n",
       "      <td id=\"T_eeca8_row28_col2\" class=\"data row28 col2\" >sag</td>\n",
       "      <td id=\"T_eeca8_row28_col3\" class=\"data row28 col3\" >auto</td>\n",
       "      <td id=\"T_eeca8_row28_col4\" class=\"data row28 col4\" >1.000000</td>\n",
       "      <td id=\"T_eeca8_row28_col5\" class=\"data row28 col5\" >0.000000</td>\n",
       "      <td id=\"T_eeca8_row28_col6\" class=\"data row28 col6\" >0.011864</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_eeca8_level0_row29\" class=\"row_heading level0 row29\" >62</th>\n",
       "      <td id=\"T_eeca8_row29_col0\" class=\"data row29 col0\" >1.000000</td>\n",
       "      <td id=\"T_eeca8_row29_col1\" class=\"data row29 col1\" >14</td>\n",
       "      <td id=\"T_eeca8_row29_col2\" class=\"data row29 col2\" >saga</td>\n",
       "      <td id=\"T_eeca8_row29_col3\" class=\"data row29 col3\" >auto</td>\n",
       "      <td id=\"T_eeca8_row29_col4\" class=\"data row29 col4\" >1.000000</td>\n",
       "      <td id=\"T_eeca8_row29_col5\" class=\"data row29 col5\" >0.000000</td>\n",
       "      <td id=\"T_eeca8_row29_col6\" class=\"data row29 col6\" >0.010468</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_eeca8_level0_row30\" class=\"row_heading level0 row30\" >34</th>\n",
       "      <td id=\"T_eeca8_row30_col0\" class=\"data row30 col0\" >0.900000</td>\n",
       "      <td id=\"T_eeca8_row30_col1\" class=\"data row30 col1\" >7</td>\n",
       "      <td id=\"T_eeca8_row30_col2\" class=\"data row30 col2\" >saga</td>\n",
       "      <td id=\"T_eeca8_row30_col3\" class=\"data row30 col3\" >auto</td>\n",
       "      <td id=\"T_eeca8_row30_col4\" class=\"data row30 col4\" >1.000000</td>\n",
       "      <td id=\"T_eeca8_row30_col5\" class=\"data row30 col5\" >0.000000</td>\n",
       "      <td id=\"T_eeca8_row30_col6\" class=\"data row30 col6\" >0.005682</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_eeca8_level0_row31\" class=\"row_heading level0 row31\" >63</th>\n",
       "      <td id=\"T_eeca8_row31_col0\" class=\"data row31 col0\" >1.000000</td>\n",
       "      <td id=\"T_eeca8_row31_col1\" class=\"data row31 col1\" >14</td>\n",
       "      <td id=\"T_eeca8_row31_col2\" class=\"data row31 col2\" >lbfgs</td>\n",
       "      <td id=\"T_eeca8_row31_col3\" class=\"data row31 col3\" >auto</td>\n",
       "      <td id=\"T_eeca8_row31_col4\" class=\"data row31 col4\" >1.000000</td>\n",
       "      <td id=\"T_eeca8_row31_col5\" class=\"data row31 col5\" >0.000000</td>\n",
       "      <td id=\"T_eeca8_row31_col6\" class=\"data row31 col6\" >0.010024</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_eeca8_level0_row32\" class=\"row_heading level0 row32\" >16</th>\n",
       "      <td id=\"T_eeca8_row32_col0\" class=\"data row32 col0\" >0.500000</td>\n",
       "      <td id=\"T_eeca8_row32_col1\" class=\"data row32 col1\" >7</td>\n",
       "      <td id=\"T_eeca8_row32_col2\" class=\"data row32 col2\" >newton-cg</td>\n",
       "      <td id=\"T_eeca8_row32_col3\" class=\"data row32 col3\" >auto</td>\n",
       "      <td id=\"T_eeca8_row32_col4\" class=\"data row32 col4\" >1.000000</td>\n",
       "      <td id=\"T_eeca8_row32_col5\" class=\"data row32 col5\" >0.000000</td>\n",
       "      <td id=\"T_eeca8_row32_col6\" class=\"data row32 col6\" >0.011745</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_eeca8_level0_row33\" class=\"row_heading level0 row33\" >23</th>\n",
       "      <td id=\"T_eeca8_row33_col0\" class=\"data row33 col0\" >0.500000</td>\n",
       "      <td id=\"T_eeca8_row33_col1\" class=\"data row33 col1\" >10</td>\n",
       "      <td id=\"T_eeca8_row33_col2\" class=\"data row33 col2\" >lbfgs</td>\n",
       "      <td id=\"T_eeca8_row33_col3\" class=\"data row33 col3\" >auto</td>\n",
       "      <td id=\"T_eeca8_row33_col4\" class=\"data row33 col4\" >1.000000</td>\n",
       "      <td id=\"T_eeca8_row33_col5\" class=\"data row33 col5\" >0.000000</td>\n",
       "      <td id=\"T_eeca8_row33_col6\" class=\"data row33 col6\" >0.007916</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_eeca8_level0_row34\" class=\"row_heading level0 row34\" >31</th>\n",
       "      <td id=\"T_eeca8_row34_col0\" class=\"data row34 col0\" >0.500000</td>\n",
       "      <td id=\"T_eeca8_row34_col1\" class=\"data row34 col1\" >14</td>\n",
       "      <td id=\"T_eeca8_row34_col2\" class=\"data row34 col2\" >lbfgs</td>\n",
       "      <td id=\"T_eeca8_row34_col3\" class=\"data row34 col3\" >auto</td>\n",
       "      <td id=\"T_eeca8_row34_col4\" class=\"data row34 col4\" >1.000000</td>\n",
       "      <td id=\"T_eeca8_row34_col5\" class=\"data row34 col5\" >0.000000</td>\n",
       "      <td id=\"T_eeca8_row34_col6\" class=\"data row34 col6\" >0.012872</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_eeca8_level0_row35\" class=\"row_heading level0 row35\" >17</th>\n",
       "      <td id=\"T_eeca8_row35_col0\" class=\"data row35 col0\" >0.500000</td>\n",
       "      <td id=\"T_eeca8_row35_col1\" class=\"data row35 col1\" >7</td>\n",
       "      <td id=\"T_eeca8_row35_col2\" class=\"data row35 col2\" >sag</td>\n",
       "      <td id=\"T_eeca8_row35_col3\" class=\"data row35 col3\" >auto</td>\n",
       "      <td id=\"T_eeca8_row35_col4\" class=\"data row35 col4\" >1.000000</td>\n",
       "      <td id=\"T_eeca8_row35_col5\" class=\"data row35 col5\" >0.000000</td>\n",
       "      <td id=\"T_eeca8_row35_col6\" class=\"data row35 col6\" >0.006020</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_eeca8_level0_row36\" class=\"row_heading level0 row36\" >18</th>\n",
       "      <td id=\"T_eeca8_row36_col0\" class=\"data row36 col0\" >0.500000</td>\n",
       "      <td id=\"T_eeca8_row36_col1\" class=\"data row36 col1\" >7</td>\n",
       "      <td id=\"T_eeca8_row36_col2\" class=\"data row36 col2\" >saga</td>\n",
       "      <td id=\"T_eeca8_row36_col3\" class=\"data row36 col3\" >auto</td>\n",
       "      <td id=\"T_eeca8_row36_col4\" class=\"data row36 col4\" >1.000000</td>\n",
       "      <td id=\"T_eeca8_row36_col5\" class=\"data row36 col5\" >0.000000</td>\n",
       "      <td id=\"T_eeca8_row36_col6\" class=\"data row36 col6\" >0.006650</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_eeca8_level0_row37\" class=\"row_heading level0 row37\" >19</th>\n",
       "      <td id=\"T_eeca8_row37_col0\" class=\"data row37 col0\" >0.500000</td>\n",
       "      <td id=\"T_eeca8_row37_col1\" class=\"data row37 col1\" >7</td>\n",
       "      <td id=\"T_eeca8_row37_col2\" class=\"data row37 col2\" >lbfgs</td>\n",
       "      <td id=\"T_eeca8_row37_col3\" class=\"data row37 col3\" >auto</td>\n",
       "      <td id=\"T_eeca8_row37_col4\" class=\"data row37 col4\" >1.000000</td>\n",
       "      <td id=\"T_eeca8_row37_col5\" class=\"data row37 col5\" >0.000000</td>\n",
       "      <td id=\"T_eeca8_row37_col6\" class=\"data row37 col6\" >0.007189</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_eeca8_level0_row38\" class=\"row_heading level0 row38\" >20</th>\n",
       "      <td id=\"T_eeca8_row38_col0\" class=\"data row38 col0\" >0.500000</td>\n",
       "      <td id=\"T_eeca8_row38_col1\" class=\"data row38 col1\" >10</td>\n",
       "      <td id=\"T_eeca8_row38_col2\" class=\"data row38 col2\" >newton-cg</td>\n",
       "      <td id=\"T_eeca8_row38_col3\" class=\"data row38 col3\" >auto</td>\n",
       "      <td id=\"T_eeca8_row38_col4\" class=\"data row38 col4\" >1.000000</td>\n",
       "      <td id=\"T_eeca8_row38_col5\" class=\"data row38 col5\" >0.000000</td>\n",
       "      <td id=\"T_eeca8_row38_col6\" class=\"data row38 col6\" >0.007691</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_eeca8_level0_row39\" class=\"row_heading level0 row39\" >22</th>\n",
       "      <td id=\"T_eeca8_row39_col0\" class=\"data row39 col0\" >0.500000</td>\n",
       "      <td id=\"T_eeca8_row39_col1\" class=\"data row39 col1\" >10</td>\n",
       "      <td id=\"T_eeca8_row39_col2\" class=\"data row39 col2\" >saga</td>\n",
       "      <td id=\"T_eeca8_row39_col3\" class=\"data row39 col3\" >auto</td>\n",
       "      <td id=\"T_eeca8_row39_col4\" class=\"data row39 col4\" >1.000000</td>\n",
       "      <td id=\"T_eeca8_row39_col5\" class=\"data row39 col5\" >0.000000</td>\n",
       "      <td id=\"T_eeca8_row39_col6\" class=\"data row39 col6\" >0.007842</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_eeca8_level0_row40\" class=\"row_heading level0 row40\" >21</th>\n",
       "      <td id=\"T_eeca8_row40_col0\" class=\"data row40 col0\" >0.500000</td>\n",
       "      <td id=\"T_eeca8_row40_col1\" class=\"data row40 col1\" >10</td>\n",
       "      <td id=\"T_eeca8_row40_col2\" class=\"data row40 col2\" >sag</td>\n",
       "      <td id=\"T_eeca8_row40_col3\" class=\"data row40 col3\" >auto</td>\n",
       "      <td id=\"T_eeca8_row40_col4\" class=\"data row40 col4\" >1.000000</td>\n",
       "      <td id=\"T_eeca8_row40_col5\" class=\"data row40 col5\" >0.000000</td>\n",
       "      <td id=\"T_eeca8_row40_col6\" class=\"data row40 col6\" >0.008627</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_eeca8_level0_row41\" class=\"row_heading level0 row41\" >24</th>\n",
       "      <td id=\"T_eeca8_row41_col0\" class=\"data row41 col0\" >0.500000</td>\n",
       "      <td id=\"T_eeca8_row41_col1\" class=\"data row41 col1\" >13</td>\n",
       "      <td id=\"T_eeca8_row41_col2\" class=\"data row41 col2\" >newton-cg</td>\n",
       "      <td id=\"T_eeca8_row41_col3\" class=\"data row41 col3\" >auto</td>\n",
       "      <td id=\"T_eeca8_row41_col4\" class=\"data row41 col4\" >1.000000</td>\n",
       "      <td id=\"T_eeca8_row41_col5\" class=\"data row41 col5\" >0.000000</td>\n",
       "      <td id=\"T_eeca8_row41_col6\" class=\"data row41 col6\" >0.007843</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_eeca8_level0_row42\" class=\"row_heading level0 row42\" >25</th>\n",
       "      <td id=\"T_eeca8_row42_col0\" class=\"data row42 col0\" >0.500000</td>\n",
       "      <td id=\"T_eeca8_row42_col1\" class=\"data row42 col1\" >13</td>\n",
       "      <td id=\"T_eeca8_row42_col2\" class=\"data row42 col2\" >sag</td>\n",
       "      <td id=\"T_eeca8_row42_col3\" class=\"data row42 col3\" >auto</td>\n",
       "      <td id=\"T_eeca8_row42_col4\" class=\"data row42 col4\" >1.000000</td>\n",
       "      <td id=\"T_eeca8_row42_col5\" class=\"data row42 col5\" >0.000000</td>\n",
       "      <td id=\"T_eeca8_row42_col6\" class=\"data row42 col6\" >0.013440</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_eeca8_level0_row43\" class=\"row_heading level0 row43\" >26</th>\n",
       "      <td id=\"T_eeca8_row43_col0\" class=\"data row43 col0\" >0.500000</td>\n",
       "      <td id=\"T_eeca8_row43_col1\" class=\"data row43 col1\" >13</td>\n",
       "      <td id=\"T_eeca8_row43_col2\" class=\"data row43 col2\" >saga</td>\n",
       "      <td id=\"T_eeca8_row43_col3\" class=\"data row43 col3\" >auto</td>\n",
       "      <td id=\"T_eeca8_row43_col4\" class=\"data row43 col4\" >1.000000</td>\n",
       "      <td id=\"T_eeca8_row43_col5\" class=\"data row43 col5\" >0.000000</td>\n",
       "      <td id=\"T_eeca8_row43_col6\" class=\"data row43 col6\" >0.008141</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_eeca8_level0_row44\" class=\"row_heading level0 row44\" >27</th>\n",
       "      <td id=\"T_eeca8_row44_col0\" class=\"data row44 col0\" >0.500000</td>\n",
       "      <td id=\"T_eeca8_row44_col1\" class=\"data row44 col1\" >13</td>\n",
       "      <td id=\"T_eeca8_row44_col2\" class=\"data row44 col2\" >lbfgs</td>\n",
       "      <td id=\"T_eeca8_row44_col3\" class=\"data row44 col3\" >auto</td>\n",
       "      <td id=\"T_eeca8_row44_col4\" class=\"data row44 col4\" >1.000000</td>\n",
       "      <td id=\"T_eeca8_row44_col5\" class=\"data row44 col5\" >0.000000</td>\n",
       "      <td id=\"T_eeca8_row44_col6\" class=\"data row44 col6\" >0.009106</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_eeca8_level0_row45\" class=\"row_heading level0 row45\" >28</th>\n",
       "      <td id=\"T_eeca8_row45_col0\" class=\"data row45 col0\" >0.500000</td>\n",
       "      <td id=\"T_eeca8_row45_col1\" class=\"data row45 col1\" >14</td>\n",
       "      <td id=\"T_eeca8_row45_col2\" class=\"data row45 col2\" >newton-cg</td>\n",
       "      <td id=\"T_eeca8_row45_col3\" class=\"data row45 col3\" >auto</td>\n",
       "      <td id=\"T_eeca8_row45_col4\" class=\"data row45 col4\" >1.000000</td>\n",
       "      <td id=\"T_eeca8_row45_col5\" class=\"data row45 col5\" >0.000000</td>\n",
       "      <td id=\"T_eeca8_row45_col6\" class=\"data row45 col6\" >0.015017</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_eeca8_level0_row46\" class=\"row_heading level0 row46\" >29</th>\n",
       "      <td id=\"T_eeca8_row46_col0\" class=\"data row46 col0\" >0.500000</td>\n",
       "      <td id=\"T_eeca8_row46_col1\" class=\"data row46 col1\" >14</td>\n",
       "      <td id=\"T_eeca8_row46_col2\" class=\"data row46 col2\" >sag</td>\n",
       "      <td id=\"T_eeca8_row46_col3\" class=\"data row46 col3\" >auto</td>\n",
       "      <td id=\"T_eeca8_row46_col4\" class=\"data row46 col4\" >1.000000</td>\n",
       "      <td id=\"T_eeca8_row46_col5\" class=\"data row46 col5\" >0.000000</td>\n",
       "      <td id=\"T_eeca8_row46_col6\" class=\"data row46 col6\" >0.011091</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_eeca8_level0_row47\" class=\"row_heading level0 row47\" >30</th>\n",
       "      <td id=\"T_eeca8_row47_col0\" class=\"data row47 col0\" >0.500000</td>\n",
       "      <td id=\"T_eeca8_row47_col1\" class=\"data row47 col1\" >14</td>\n",
       "      <td id=\"T_eeca8_row47_col2\" class=\"data row47 col2\" >saga</td>\n",
       "      <td id=\"T_eeca8_row47_col3\" class=\"data row47 col3\" >auto</td>\n",
       "      <td id=\"T_eeca8_row47_col4\" class=\"data row47 col4\" >1.000000</td>\n",
       "      <td id=\"T_eeca8_row47_col5\" class=\"data row47 col5\" >0.000000</td>\n",
       "      <td id=\"T_eeca8_row47_col6\" class=\"data row47 col6\" >0.009661</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_eeca8_level0_row48\" class=\"row_heading level0 row48\" >8</th>\n",
       "      <td id=\"T_eeca8_row48_col0\" class=\"data row48 col0\" >0.001000</td>\n",
       "      <td id=\"T_eeca8_row48_col1\" class=\"data row48 col1\" >13</td>\n",
       "      <td id=\"T_eeca8_row48_col2\" class=\"data row48 col2\" >newton-cg</td>\n",
       "      <td id=\"T_eeca8_row48_col3\" class=\"data row48 col3\" >auto</td>\n",
       "      <td id=\"T_eeca8_row48_col4\" class=\"data row48 col4\" >0.998780</td>\n",
       "      <td id=\"T_eeca8_row48_col5\" class=\"data row48 col5\" >0.001220</td>\n",
       "      <td id=\"T_eeca8_row48_col6\" class=\"data row48 col6\" >0.029229</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_eeca8_level0_row49\" class=\"row_heading level0 row49\" >2</th>\n",
       "      <td id=\"T_eeca8_row49_col0\" class=\"data row49 col0\" >0.001000</td>\n",
       "      <td id=\"T_eeca8_row49_col1\" class=\"data row49 col1\" >7</td>\n",
       "      <td id=\"T_eeca8_row49_col2\" class=\"data row49 col2\" >saga</td>\n",
       "      <td id=\"T_eeca8_row49_col3\" class=\"data row49 col3\" >auto</td>\n",
       "      <td id=\"T_eeca8_row49_col4\" class=\"data row49 col4\" >0.998780</td>\n",
       "      <td id=\"T_eeca8_row49_col5\" class=\"data row49 col5\" >0.001220</td>\n",
       "      <td id=\"T_eeca8_row49_col6\" class=\"data row49 col6\" >0.006349</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_eeca8_level0_row50\" class=\"row_heading level0 row50\" >3</th>\n",
       "      <td id=\"T_eeca8_row50_col0\" class=\"data row50 col0\" >0.001000</td>\n",
       "      <td id=\"T_eeca8_row50_col1\" class=\"data row50 col1\" >7</td>\n",
       "      <td id=\"T_eeca8_row50_col2\" class=\"data row50 col2\" >lbfgs</td>\n",
       "      <td id=\"T_eeca8_row50_col3\" class=\"data row50 col3\" >auto</td>\n",
       "      <td id=\"T_eeca8_row50_col4\" class=\"data row50 col4\" >0.998780</td>\n",
       "      <td id=\"T_eeca8_row50_col5\" class=\"data row50 col5\" >0.001220</td>\n",
       "      <td id=\"T_eeca8_row50_col6\" class=\"data row50 col6\" >0.006760</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_eeca8_level0_row51\" class=\"row_heading level0 row51\" >4</th>\n",
       "      <td id=\"T_eeca8_row51_col0\" class=\"data row51 col0\" >0.001000</td>\n",
       "      <td id=\"T_eeca8_row51_col1\" class=\"data row51 col1\" >10</td>\n",
       "      <td id=\"T_eeca8_row51_col2\" class=\"data row51 col2\" >newton-cg</td>\n",
       "      <td id=\"T_eeca8_row51_col3\" class=\"data row51 col3\" >auto</td>\n",
       "      <td id=\"T_eeca8_row51_col4\" class=\"data row51 col4\" >0.998780</td>\n",
       "      <td id=\"T_eeca8_row51_col5\" class=\"data row51 col5\" >0.001220</td>\n",
       "      <td id=\"T_eeca8_row51_col6\" class=\"data row51 col6\" >0.007533</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_eeca8_level0_row52\" class=\"row_heading level0 row52\" >5</th>\n",
       "      <td id=\"T_eeca8_row52_col0\" class=\"data row52 col0\" >0.001000</td>\n",
       "      <td id=\"T_eeca8_row52_col1\" class=\"data row52 col1\" >10</td>\n",
       "      <td id=\"T_eeca8_row52_col2\" class=\"data row52 col2\" >sag</td>\n",
       "      <td id=\"T_eeca8_row52_col3\" class=\"data row52 col3\" >auto</td>\n",
       "      <td id=\"T_eeca8_row52_col4\" class=\"data row52 col4\" >0.998780</td>\n",
       "      <td id=\"T_eeca8_row52_col5\" class=\"data row52 col5\" >0.001220</td>\n",
       "      <td id=\"T_eeca8_row52_col6\" class=\"data row52 col6\" >0.007277</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_eeca8_level0_row53\" class=\"row_heading level0 row53\" >6</th>\n",
       "      <td id=\"T_eeca8_row53_col0\" class=\"data row53 col0\" >0.001000</td>\n",
       "      <td id=\"T_eeca8_row53_col1\" class=\"data row53 col1\" >10</td>\n",
       "      <td id=\"T_eeca8_row53_col2\" class=\"data row53 col2\" >saga</td>\n",
       "      <td id=\"T_eeca8_row53_col3\" class=\"data row53 col3\" >auto</td>\n",
       "      <td id=\"T_eeca8_row53_col4\" class=\"data row53 col4\" >0.998780</td>\n",
       "      <td id=\"T_eeca8_row53_col5\" class=\"data row53 col5\" >0.001220</td>\n",
       "      <td id=\"T_eeca8_row53_col6\" class=\"data row53 col6\" >0.008485</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_eeca8_level0_row54\" class=\"row_heading level0 row54\" >7</th>\n",
       "      <td id=\"T_eeca8_row54_col0\" class=\"data row54 col0\" >0.001000</td>\n",
       "      <td id=\"T_eeca8_row54_col1\" class=\"data row54 col1\" >10</td>\n",
       "      <td id=\"T_eeca8_row54_col2\" class=\"data row54 col2\" >lbfgs</td>\n",
       "      <td id=\"T_eeca8_row54_col3\" class=\"data row54 col3\" >auto</td>\n",
       "      <td id=\"T_eeca8_row54_col4\" class=\"data row54 col4\" >0.998780</td>\n",
       "      <td id=\"T_eeca8_row54_col5\" class=\"data row54 col5\" >0.001220</td>\n",
       "      <td id=\"T_eeca8_row54_col6\" class=\"data row54 col6\" >0.010671</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_eeca8_level0_row55\" class=\"row_heading level0 row55\" >1</th>\n",
       "      <td id=\"T_eeca8_row55_col0\" class=\"data row55 col0\" >0.001000</td>\n",
       "      <td id=\"T_eeca8_row55_col1\" class=\"data row55 col1\" >7</td>\n",
       "      <td id=\"T_eeca8_row55_col2\" class=\"data row55 col2\" >sag</td>\n",
       "      <td id=\"T_eeca8_row55_col3\" class=\"data row55 col3\" >auto</td>\n",
       "      <td id=\"T_eeca8_row55_col4\" class=\"data row55 col4\" >0.998780</td>\n",
       "      <td id=\"T_eeca8_row55_col5\" class=\"data row55 col5\" >0.001220</td>\n",
       "      <td id=\"T_eeca8_row55_col6\" class=\"data row55 col6\" >0.006425</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_eeca8_level0_row56\" class=\"row_heading level0 row56\" >9</th>\n",
       "      <td id=\"T_eeca8_row56_col0\" class=\"data row56 col0\" >0.001000</td>\n",
       "      <td id=\"T_eeca8_row56_col1\" class=\"data row56 col1\" >13</td>\n",
       "      <td id=\"T_eeca8_row56_col2\" class=\"data row56 col2\" >sag</td>\n",
       "      <td id=\"T_eeca8_row56_col3\" class=\"data row56 col3\" >auto</td>\n",
       "      <td id=\"T_eeca8_row56_col4\" class=\"data row56 col4\" >0.998780</td>\n",
       "      <td id=\"T_eeca8_row56_col5\" class=\"data row56 col5\" >0.001220</td>\n",
       "      <td id=\"T_eeca8_row56_col6\" class=\"data row56 col6\" >0.008225</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_eeca8_level0_row57\" class=\"row_heading level0 row57\" >10</th>\n",
       "      <td id=\"T_eeca8_row57_col0\" class=\"data row57 col0\" >0.001000</td>\n",
       "      <td id=\"T_eeca8_row57_col1\" class=\"data row57 col1\" >13</td>\n",
       "      <td id=\"T_eeca8_row57_col2\" class=\"data row57 col2\" >saga</td>\n",
       "      <td id=\"T_eeca8_row57_col3\" class=\"data row57 col3\" >auto</td>\n",
       "      <td id=\"T_eeca8_row57_col4\" class=\"data row57 col4\" >0.998780</td>\n",
       "      <td id=\"T_eeca8_row57_col5\" class=\"data row57 col5\" >0.001220</td>\n",
       "      <td id=\"T_eeca8_row57_col6\" class=\"data row57 col6\" >0.008071</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_eeca8_level0_row58\" class=\"row_heading level0 row58\" >11</th>\n",
       "      <td id=\"T_eeca8_row58_col0\" class=\"data row58 col0\" >0.001000</td>\n",
       "      <td id=\"T_eeca8_row58_col1\" class=\"data row58 col1\" >13</td>\n",
       "      <td id=\"T_eeca8_row58_col2\" class=\"data row58 col2\" >lbfgs</td>\n",
       "      <td id=\"T_eeca8_row58_col3\" class=\"data row58 col3\" >auto</td>\n",
       "      <td id=\"T_eeca8_row58_col4\" class=\"data row58 col4\" >0.998780</td>\n",
       "      <td id=\"T_eeca8_row58_col5\" class=\"data row58 col5\" >0.001220</td>\n",
       "      <td id=\"T_eeca8_row58_col6\" class=\"data row58 col6\" >0.008297</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_eeca8_level0_row59\" class=\"row_heading level0 row59\" >12</th>\n",
       "      <td id=\"T_eeca8_row59_col0\" class=\"data row59 col0\" >0.001000</td>\n",
       "      <td id=\"T_eeca8_row59_col1\" class=\"data row59 col1\" >14</td>\n",
       "      <td id=\"T_eeca8_row59_col2\" class=\"data row59 col2\" >newton-cg</td>\n",
       "      <td id=\"T_eeca8_row59_col3\" class=\"data row59 col3\" >auto</td>\n",
       "      <td id=\"T_eeca8_row59_col4\" class=\"data row59 col4\" >0.998780</td>\n",
       "      <td id=\"T_eeca8_row59_col5\" class=\"data row59 col5\" >0.001220</td>\n",
       "      <td id=\"T_eeca8_row59_col6\" class=\"data row59 col6\" >0.009978</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_eeca8_level0_row60\" class=\"row_heading level0 row60\" >13</th>\n",
       "      <td id=\"T_eeca8_row60_col0\" class=\"data row60 col0\" >0.001000</td>\n",
       "      <td id=\"T_eeca8_row60_col1\" class=\"data row60 col1\" >14</td>\n",
       "      <td id=\"T_eeca8_row60_col2\" class=\"data row60 col2\" >sag</td>\n",
       "      <td id=\"T_eeca8_row60_col3\" class=\"data row60 col3\" >auto</td>\n",
       "      <td id=\"T_eeca8_row60_col4\" class=\"data row60 col4\" >0.998780</td>\n",
       "      <td id=\"T_eeca8_row60_col5\" class=\"data row60 col5\" >0.001220</td>\n",
       "      <td id=\"T_eeca8_row60_col6\" class=\"data row60 col6\" >0.009400</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_eeca8_level0_row61\" class=\"row_heading level0 row61\" >14</th>\n",
       "      <td id=\"T_eeca8_row61_col0\" class=\"data row61 col0\" >0.001000</td>\n",
       "      <td id=\"T_eeca8_row61_col1\" class=\"data row61 col1\" >14</td>\n",
       "      <td id=\"T_eeca8_row61_col2\" class=\"data row61 col2\" >saga</td>\n",
       "      <td id=\"T_eeca8_row61_col3\" class=\"data row61 col3\" >auto</td>\n",
       "      <td id=\"T_eeca8_row61_col4\" class=\"data row61 col4\" >0.998780</td>\n",
       "      <td id=\"T_eeca8_row61_col5\" class=\"data row61 col5\" >0.001220</td>\n",
       "      <td id=\"T_eeca8_row61_col6\" class=\"data row61 col6\" >0.011705</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_eeca8_level0_row62\" class=\"row_heading level0 row62\" >15</th>\n",
       "      <td id=\"T_eeca8_row62_col0\" class=\"data row62 col0\" >0.001000</td>\n",
       "      <td id=\"T_eeca8_row62_col1\" class=\"data row62 col1\" >14</td>\n",
       "      <td id=\"T_eeca8_row62_col2\" class=\"data row62 col2\" >lbfgs</td>\n",
       "      <td id=\"T_eeca8_row62_col3\" class=\"data row62 col3\" >auto</td>\n",
       "      <td id=\"T_eeca8_row62_col4\" class=\"data row62 col4\" >0.998780</td>\n",
       "      <td id=\"T_eeca8_row62_col5\" class=\"data row62 col5\" >0.001220</td>\n",
       "      <td id=\"T_eeca8_row62_col6\" class=\"data row62 col6\" >0.012531</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_eeca8_level0_row63\" class=\"row_heading level0 row63\" >0</th>\n",
       "      <td id=\"T_eeca8_row63_col0\" class=\"data row63 col0\" >0.001000</td>\n",
       "      <td id=\"T_eeca8_row63_col1\" class=\"data row63 col1\" >7</td>\n",
       "      <td id=\"T_eeca8_row63_col2\" class=\"data row63 col2\" >newton-cg</td>\n",
       "      <td id=\"T_eeca8_row63_col3\" class=\"data row63 col3\" >auto</td>\n",
       "      <td id=\"T_eeca8_row63_col4\" class=\"data row63 col4\" >0.998780</td>\n",
       "      <td id=\"T_eeca8_row63_col5\" class=\"data row63 col5\" >0.001220</td>\n",
       "      <td id=\"T_eeca8_row63_col6\" class=\"data row63 col6\" >0.007117</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n"
      ],
      "text/plain": [
       "<pandas.io.formats.style.Styler at 0x7fa566580fd0>"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Logit\n",
    "\n",
    "index = 42\n",
    "logit_results.style.background_gradient(\n",
    "    subset=[\"n_bits\", \"mean_test_score\", \"fhe_inference_time\"]\n",
    ").apply(\n",
    "    lambda index: [\"background-color: green\"] * len(index), axis=1, subset=pd.IndexSlice[[index], :]\n",
    ")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "In real-time applications, inference time is a crucial factor.\n",
    "\n",
    "From XGB's results, we can see that most accurate models are also the ones that have the longest inference times in FHE. However, considering less complex configurations could can offer a viable compromise between efficiency and accuracy.\n",
    "\n",
    "In this specific use case, Logistic Regression outperforms XGB in terms of both accuracy and time computing."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Most optimal model in FHE "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": [
    "# The most optimal model in the FHE realm is:\n",
    "\n",
    "optimal_param = {\"C\": 0.9, \"n_bits\": 13, \"solver\": \"sag\", \"multi_class\": \"auto\"}\n",
    "\n",
    "clf = ConcreteLogisticRegression(**optimal_param)\n",
    "\n",
    "clf.fit(X_train, y_train)\n",
    "\n",
    "fhe_circuit = clf.compile(X_train)\n",
    "\n",
    "fhe_circuit.client.keygen(force=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Inference timing: 0.003 sPrediction: 30 -> Migraine\n"
     ]
    }
   ],
   "source": [
    "# Simulated FHE mode\n",
    "\n",
    "start_time = time()\n",
    "y_pred = clf.predict(random_x, fhe=\"simulate\")\n",
    "\n",
    "print(\n",
    "    f\"Inference timing: {time() - start_time:.3f} s\"\n",
    "    f\"Prediction: {y_pred[0]} -> \"\n",
    "    f\"{df_test[df_test['prognosis_encoded'] == y_pred[0]]['prognosis'].values[0]}\"\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Inference timing: 0.014 sPrediction: 30 -> Migraine\n"
     ]
    }
   ],
   "source": [
    "# FHE mode\n",
    "\n",
    "start_time = time()\n",
    "y_pred = clf.predict(random_x, fhe=\"execute\")\n",
    "\n",
    "print(\n",
    "    f\"Inference timing: {time() - start_time:.3f} s\"\n",
    "    f\"Prediction: {y_pred[0]} -> \"\n",
    "    f\"{df_test[df_test['prognosis_encoded'] == y_pred[0]]['prognosis'].values[0]}\"\n",
    ")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Save the model for deployment using Concrete ML"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Now that we have selected the optimal model, we can save it."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "from concrete.ml.deployment import FHEModelDev\n",
    "\n",
    "path_to_model = Path(\"./health_prediction/deployment\").resolve()\n",
    "\n",
    "if path_to_model.exists():\n",
    "    shutil.rmtree(path_to_model)\n",
    "\n",
    "dev = FHEModelDev(path_to_model, clf)\n",
    "dev.save(via_mlir=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Conclusion\n",
    "\n",
    "This notebook shows how to carefully choose a Concrete ML model for deployment by highlighting the subtleties between the most accurate and the optimal model within FHE constraints.\n",
    "\n",
    "By leveraging fully homomorphic encryption (FHE), it becomes possible to process sensitive data, such as patient symptoms and their diagnosis, while preserving individuals' privacy and enabling advancements in the healthcare domain."
   ]
  }
 ],
 "metadata": {
  "execution": {
   "timeout": 10800
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}