Practicing_Machine_learning_for_cancer_prediction.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "dbd1ba08-8c33-49f5-99d9-ae4f97f0e07e",
   "metadata": {},
   "source": [
    "# Here, I learn how to use RNA-seq gene expression data for cancer types using Vappiah's tutorial accessible on Youtube and Github"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "000d8431-0531-4db1-a82b-6f57adb661b2",
   "metadata": {},
   "source": [
    "## The goal is to practice using Jupyter notebook, learn Python, and data analysis relevant for Bioinformatics"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "7c25091f-a8e8-4b8e-95ce-5596fd07a2b7",
   "metadata": {},
   "source": [
    "### Notebook by G. Yagudayeva, November, 2024"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "6f3a9209-a267-4d10-85c8-7a310da3afe1",
   "metadata": {},
   "source": [
   ]
  },
  {
   "cell_type": "markdown",
   "id": "7b57cb54-882e-4bb7-b551-463e8a742279",
   "metadata": {},
   "source": [
    "### Required libraries are: numpy, matplotlib, seaborn, pandas, and scikit. numpy and pandas are for data handling; matplotlib and seaborn are for data visualization, and scikit are for preprocessing, feature selection, classification, and performance metrics"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "354edd77-796d-4d5a-a06c-5ca49dc00dc0",
   "metadata": {},
   "outputs": [],
   "source": [
    "#Let's call these libraries"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "806f358e-8ea9-40b8-9564-b6568941704a",
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.preprocessing import LabelEncoder\n",
    "from sklearn.preprocessing import label_binarize\n",
    "from sklearn.preprocessing import MinMaxScaler\n",
    "from sklearn.feature_selection import mutual_info_classif\n",
    "from sklearn.multiclass import OneVsRestClassifier\n",
    "from sklearn.ensemble import RandomForestClassifier\n",
    "from sklearn.metrics import balanced_accuracy_score,f1_score,precision_score, recall_score\n",
    "from sklearn.metrics import classification_report\n",
    "from sklearn.metrics import confusion_matrix\n",
    "from sklearn.metrics import plot_confusion_matrix\n",
    "from sklearn.metrics import roc_curve,auc\n",
    "from sklearn.metrics import roc_auc_score"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "bc37194f-5a4b-498d-9885-cff2c5c5eed8",
   "metadata": {},
   "outputs": [],
   "source": [
    "#Now, let's get the data we will use from a git repo using Vappiah's data for the tutorial"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "id": "c4712a24-1a13-48f8-9dcc-9190b5146027",
   "metadata": {},
   "outputs": [],
   "source": [
    "file_url='https://github.com/vappiah/Machine-Learning-Tutorials/raw/main/datasets/cancer_gene_expression.zip'\n",
    "\n",
    "dataframe=pd.read_csv(file_url)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "5818f27d-ec32-45a6-a951-19e549993178",
   "metadata": {},
   "outputs": [],
   "source": [
    "#Let's look at the data we just got, what does it look like? Do we need to clean it?"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "5f643662-6c55-4588-8e42-6a2ba318beec",
   "metadata": {},
   "outputs": [],
   "source": [
    "#We can see the number of samples and features"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "id": "cce15465-680a-46c6-a42e-3aa70fc10ee4",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(801, 8001)\n"
     ]
    }
   ],
   "source": [
    "print(dataframe.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "2bb4a1ab-b12a-49fb-85f5-9a6c8204dc13",
   "metadata": {},
   "outputs": [],
   "source": [
    "#Now, we can see what the columns look like"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "id": "d5748926-1c0c-445f-aa7b-57f0729f4312",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Index(['gene_1', 'gene_2', 'gene_3'], dtype='object')\n"
     ]
    }
   ],
   "source": [
    "print(dataframe.columns[0:3])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "7ef55bda-9ada-466f-93d3-3ba8810036bc",
   "metadata": {},
   "outputs": [],
   "source": [
    "#How about the last column of this data frame?"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "id": "2173120e-ae24-467e-a52e-7c6d2d0842fa",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'Cancer_Type'"
      ]
     },
     "execution_count": 25,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "dataframe.columns[-1]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "0beb06e7-804c-4d7b-bb76-bed77e48d813",
   "metadata": {},
   "outputs": [],
   "source": [
    "#What values are missing?"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "id": "f3c543f4-d7db-4a5d-a356-3029a7f62528",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "columns with missing values:0\n"
     ]
    }
   ],
   "source": [
    "datanul=dataframe.isnull().sum()\n",
    "g=[i for i in datanul if i>0]\n",
    "\n",
    "print('columns with missing values:%d'%len(g))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "50bf2c3c-fb93-40e6-b52c-8d3d7da71bc1",
   "metadata": {},
   "outputs": [],
   "source": [
    "#How many cancer types are there? or classes?"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "id": "92a41087-07ea-4e46-96b5-62ab81009a08",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Cancer_Type\n",
      "BRCA    300\n",
      "KIRC    146\n",
      "LUAD    141\n",
      "PRAD    136\n",
      "COAD     78\n",
      "Name: count, dtype: int64\n"
     ]
    }
   ],
   "source": [
    "print(dataframe['Cancer_Type'].value_counts())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "f27b5638-6247-417c-968e-df43d9d91a4d",
   "metadata": {},
   "outputs": [],
   "source": [
    "#Let's do some visualization, can we see a bar chart to display the class distribution?"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "id": "e8fb64d2-e347-4566-ae0f-393ee48d4004",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<Axes: xlabel='Cancer_Type'>"
      ]
     },
     "execution_count": 31,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "dataframe['Cancer_Type'].value_counts().plot.bar()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "6cc56fcc-fee9-4b84-8f34-0a0f6e8f77cd",
   "metadata": {},
   "source": [
    "### Awesome!! We were able to do a nice visualization for what the class distribution would look like in a bar chart"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "e188c50f-b639-4626-b6af-9cce8aaec7fd",
   "metadata": {},
   "outputs": [],
   "source": [
    "#Now, we need to pre-process our data, and put it into the correct format before modelling so we can look at the data besst and most accurately"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "0c1d838a-e37f-41b5-87dd-b7d0b6036418",
   "metadata": {},
   "outputs": [],
   "source": [
    "#To start, we can separate feature values from the class to satisfy scikit"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "id": "dac747d4-f41a-4c12-ba8a-d85624558f90",
   "metadata": {},
   "outputs": [],
   "source": [
    "X=dataframe.iloc[:,0:-1]\n",
    "y=dataframe.iloc[:,-1]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "id": "4824eb5d-d139-4a57-87ab-1909ee643420",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(801, 8000)"
      ]
     },
     "execution_count": 37,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "id": "91f7cb74-d468-459f-b9ca-f0849afdd75c",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(801,)"
      ]
     },
     "execution_count": 39,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "c1ac25b8-b1f2-4881-89d5-7bf0c3644a20",
   "metadata": {},
   "outputs": [],
   "source": [
    "#When we work with machine learning (ML), we need to encode labels for our data, what does this mean? Well, ML models require input data to be in numeric forms, so we can encode and convert them to be numeric forms of data to conduct future analyses"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "7e7e8f53-4dc3-4e3b-9666-edcac4eb767c",
   "metadata": {},
   "outputs": [],
   "source": [
    "#To encode values between 0 and n classes-1 using the labelencoder tool"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "id": "f566a14c-eb0a-4f12-875f-837187e63185",
   "metadata": {
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "label_encoder=LabelEncoder()\n",
    "label_encoder.fit(y)\n",
    "y_encoded=label_encoder.transform(y)\n",
    "labels=label_encoder.classes_\n",
    "classes=np.unique(y_encoded)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "id": "93fb3496-214e-4814-8db9-632ab4173b14",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array(['BRCA', 'COAD', 'KIRC', 'LUAD', 'PRAD'], dtype=object)"
      ]
     },
     "execution_count": 43,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "labels"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "id": "15d3b012-0153-4c3d-a627-fa47d9797bd1",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([0, 1, 2, 3, 4])"
      ]
     },
     "execution_count": 45,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "classes"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "id": "0580365a-5043-499d-8d75-68d63ddd21b7",
   "metadata": {},
   "outputs": [],
   "source": [
    "#Now, we need to split the data into training and test subsets, because we need the ML to be able to identify iscriminatory paterns to make future predictions. Usually when ML models are generated, they need to be tested on practice data before we use real data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "4111508e-8f7e-46b7-a273-e099929f46cd",
   "metadata": {},
   "outputs": [],
   "source": [
    "#split data into training and test sets"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "id": "fcbfe063-0535-44fc-9599-dad2da6a58b2",
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train,X_test,y_train,y_test=train_test_split(X,y_encoded,test_size=0.2,random_state=42)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "20024067-6335-4d81-ad98-370017da554a",
   "metadata": {},
   "outputs": [],
   "source": [
    "#Now, let's visualize it in the data frame:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "id": "09885025-e18d-4e36-97f1-2a3f33bd722f",
   "metadata": {
    "scrolled": true
   },
   "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>gene_1</th>\n",
       "      <th>gene_2</th>\n",
       "      <th>gene_3</th>\n",
       "      <th>gene_4</th>\n",
       "      <th>gene_5</th>\n",
       "      <th>gene_6</th>\n",
       "      <th>gene_7</th>\n",
       "      <th>gene_8</th>\n",
       "      <th>gene_9</th>\n",
       "      <th>gene_10</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>801.000000</td>\n",
       "      <td>801.000000</td>\n",
       "      <td>801.000000</td>\n",
       "      <td>801.000000</td>\n",
       "      <td>801.000000</td>\n",
       "      <td>801.000000</td>\n",
       "      <td>801.000000</td>\n",
       "      <td>801.000000</td>\n",
       "      <td>801.000000</td>\n",
       "      <td>801.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>0.026642</td>\n",
       "      <td>0.231049</td>\n",
       "      <td>0.010844</td>\n",
       "      <td>2.405459</td>\n",
       "      <td>0.005799</td>\n",
       "      <td>0.065640</td>\n",
       "      <td>0.022036</td>\n",
       "      <td>0.028394</td>\n",
       "      <td>0.703998</td>\n",
       "      <td>1.916364</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>0.136850</td>\n",
       "      <td>0.656692</td>\n",
       "      <td>0.153859</td>\n",
       "      <td>0.797236</td>\n",
       "      <td>0.052265</td>\n",
       "      <td>0.213557</td>\n",
       "      <td>0.148917</td>\n",
       "      <td>0.161606</td>\n",
       "      <td>0.779321</td>\n",
       "      <td>0.877010</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25%</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>1.898595</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>1.357214</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50%</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>2.358284</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.518031</td>\n",
       "      <td>1.877862</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>2.929355</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>1.123931</td>\n",
       "      <td>2.486792</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>1.482332</td>\n",
       "      <td>6.575004</td>\n",
       "      <td>3.778061</td>\n",
       "      <td>4.746646</td>\n",
       "      <td>0.570754</td>\n",
       "      <td>1.866275</td>\n",
       "      <td>2.169925</td>\n",
       "      <td>2.375957</td>\n",
       "      <td>4.541670</td>\n",
       "      <td>4.932231</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "           gene_1      gene_2      gene_3      gene_4      gene_5      gene_6  \\\n",
       "count  801.000000  801.000000  801.000000  801.000000  801.000000  801.000000   \n",
       "mean     0.026642    0.231049    0.010844    2.405459    0.005799    0.065640   \n",
       "std      0.136850    0.656692    0.153859    0.797236    0.052265    0.213557   \n",
       "min      0.000000    0.000000    0.000000    0.000000    0.000000    0.000000   \n",
       "25%      0.000000    0.000000    0.000000    1.898595    0.000000    0.000000   \n",
       "50%      0.000000    0.000000    0.000000    2.358284    0.000000    0.000000   \n",
       "75%      0.000000    0.000000    0.000000    2.929355    0.000000    0.000000   \n",
       "max      1.482332    6.575004    3.778061    4.746646    0.570754    1.866275   \n",
       "\n",
       "           gene_7      gene_8      gene_9     gene_10  \n",
       "count  801.000000  801.000000  801.000000  801.000000  \n",
       "mean     0.022036    0.028394    0.703998    1.916364  \n",
       "std      0.148917    0.161606    0.779321    0.877010  \n",
       "min      0.000000    0.000000    0.000000    0.000000  \n",
       "25%      0.000000    0.000000    0.000000    1.357214  \n",
       "50%      0.000000    0.000000    0.518031    1.877862  \n",
       "75%      0.000000    0.000000    1.123931    2.486792  \n",
       "max      2.169925    2.375957    4.541670    4.932231  "
      ]
     },
     "execution_count": 52,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "dataframe.iloc[:,0:10].describe()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "a124be2e-d8c4-4bef-91a6-a536cc2b1c98",
   "metadata": {},
   "outputs": [],
   "source": [
    "#Data normalization must be done to normalize the data to be within the same range and improve bias and ML model performance"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "f508f7d1-ce6e-40ba-a74a-a36f8982f1f0",
   "metadata": {},
   "outputs": [],
   "source": [
    "#Let's scale the data between 0 and 1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "id": "13b75c35-0479-481a-b581-9beb5d2c5834",
   "metadata": {},
   "outputs": [],
   "source": [
    "min_max_scaler=MinMaxScaler()\n",
    "X_train_norm=min_max_scaler.fit_transform(X_train)\n",
    "X_test_norm=min_max_scaler.fit_transform(X_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "691b9cc9-fb68-4b92-949c-672e4c13d3df",
   "metadata": {},
   "outputs": [],
   "source": [
    "#Before you do ML, another preprocessing step to it is to do feature selection; feature selection is a method to pre-process the data and select relevant featurs for classiciation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "416d6ebc-7055-4171-81bf-646aeea08d15",
   "metadata": {},
   "outputs": [],
   "source": [
    "#For our purposes, let's do mutual information algorithm to compute the relevance of each feature with the stop n features to select for our future ML analyses"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "af079772-3a36-49bb-8a53-b6252cdbd42c",
   "metadata": {},
   "outputs": [],
   "source": [
    "#We can do feature selection from mutual information"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "id": "64fa69dd-4321-4b22-9b7e-f5253ce7075a",
   "metadata": {},
   "outputs": [],
   "source": [
    "MI=mutual_info_classif(X_train_norm,y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "da1adace-f480-47f3-951c-c30f57ebc5e8",
   "metadata": {},
   "outputs": [],
   "source": [
    "#We can select the top n features and let's try to do it 300 (you can modify this number and see how the ML performance changes overtime"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "id": "f908b2d4-efa2-4260-8941-8ad6110a00fa",
   "metadata": {},
   "outputs": [],
   "source": [
    "n_features=300\n",
    "selected_scores_indices=np.argsort(MI)[::-1][0:n_features]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 60,
   "id": "316d98fa-5c65-4c44-a2e0-7b84f7764093",
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train_selected=X_train_norm[:,selected_scores_indices]\n",
    "X_test_selected=X_test_norm[:,selected_scores_indices]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 62,
   "id": "0489436c-b715-46f5-91bb-f4279f81e93f",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(640, 300)"
      ]
     },
     "execution_count": 62,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_train_selected.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 64,
   "id": "2cfe1320-0dc0-4148-8390-e6a02639fe79",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(161, 300)"
      ]
     },
     "execution_count": 64,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_test_selected.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "150af1c5-bd39-46fd-88e6-b9b8a0d8e535",
   "metadata": {},
   "outputs": [],
   "source": [
    "#Now we will do the classification, which is the training and testing of the ML model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "d649ca9f-bb17-4119-a236-aa5759580c94",
   "metadata": {},
   "outputs": [],
   "source": [
    "#Here, we will use the random forest classifier, which can be used for multiclass and high dimensional data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "ef28a812-833e-4047-89ad-1eb13a3372f6",
   "metadata": {},
   "outputs": [],
   "source": [
    "#Via model training, we can allow the ML model to learn from the data and identify similar patterns to predict the outcomes of the data it has never sen before; essentially, we are articficially training the model to detect specific patterns"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "e40d257d-a9e9-4204-9f10-e74d5eda47b5",
   "metadata": {},
   "outputs": [],
   "source": [
    "#Let's use random forest classifier and teach the model to predict each class against the other"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 66,
   "id": "0029f17b-2c36-4634-95ed-db961bd7bfcc",
   "metadata": {},
   "outputs": [],
   "source": [
    "RF=OneVsRestClassifier(RandomForestClassifier(max_features=0.2))\n",
    "RF.fit(X_train_selected,y_train)\n",
    "y_pred =RF.predict(X_test_selected)\n",
    "pred_prob = RF.predict_proba(X_test_selected)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "a3181d1a-ce87-4cda-a57f-906c1d405e58",
   "metadata": {},
   "outputs": [],
   "source": [
    "#Now, how well did we train out model? we need to evalute the model and see how good it learned the artifiical intelligence we gave it"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "f444ca22-99f6-45b7-a8d6-ab043ffdc0e5",
   "metadata": {},
   "outputs": [],
   "source": [
    "#We will test it using the following metrics: accuracy, precision, recall, f1 score, confusion matrix, and ROC curve"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "dcce405b-5dc0-4a9b-b8e9-41bbe2e5d48b",
   "metadata": {},
   "outputs": [],
   "source": [
    "#Let's test all these and see what we get"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 68,
   "id": "d0123bc2-e738-4704-b6fa-2820f5015b29",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "accuracy:0.9708\n",
      "precision:0.9815\n",
      "recall:0.9814\n",
      "f1score:0.9812\n",
      "\n",
      "\n",
      "classification report\n",
      "\n",
      "\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "        BRCA       0.97      1.00      0.98        60\n",
      "        COAD       1.00      0.94      0.97        16\n",
      "        KIRC       1.00      1.00      1.00        28\n",
      "        LUAD       0.96      0.92      0.94        24\n",
      "        PRAD       1.00      1.00      1.00        33\n",
      "\n",
      "    accuracy                           0.98       161\n",
      "   macro avg       0.98      0.97      0.98       161\n",
      "weighted avg       0.98      0.98      0.98       161\n",
      "\n"
     ]
    }
   ],
   "source": [
    "accuracy=np.round(balanced_accuracy_score(y_test,y_pred),4)\n",
    "print('accuracy:%0.4f'%accuracy)\n",
    "precision=np.round(precision_score(y_test,y_pred,average = 'weighted'),4)\n",
    "print('precision:%0.4f'%precision)\n",
    "recall=np.round(recall_score(y_test,y_pred,average = 'weighted'),4)\n",
    "print('recall:%0.4f'%recall)\n",
    "f1score=np.round(f1_score(y_test,y_pred,average = 'weighted'),4)\n",
    "print('f1score:%0.4f'%f1score)\n",
    "report=classification_report(y_test,y_pred, target_names=labels)\n",
    "print('\\n')\n",
    "print('classification report\\n\\n')\n",
    "print(report)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "d58d9b38-d396-4316-9cf6-ac091c05e25c",
   "metadata": {},
   "outputs": [],
   "source": [
    "#Now, separate for the confusion matrix (visualized by seaborn)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 80,
   "id": "0e6945eb-3753-4b31-affa-39665afb4297",
   "metadata": {},
   "outputs": [],
   "source": [
    "cm=confusion_matrix(y_test,y_pred)\n",
    "cm_df=pd.DataFrame(cm,index=labels,columns=labels)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 78,
   "id": "d00d0736-9b00-49d3-9754-5bc7e3be0a7c",
   "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>BRCA</th>\n",
       "      <th>COAD</th>\n",
       "      <th>KIRC</th>\n",
       "      <th>LUAD</th>\n",
       "      <th>PRAD</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>BRCA</th>\n",
       "      <td>60</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>COAD</th>\n",
       "      <td>0</td>\n",
       "      <td>15</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>KIRC</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>28</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>LUAD</th>\n",
       "      <td>2</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>22</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>PRAD</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>33</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "      BRCA  COAD  KIRC  LUAD  PRAD\n",
       "BRCA    60     0     0     0     0\n",
       "COAD     0    15     0     1     0\n",
       "KIRC     0     0    28     0     0\n",
       "LUAD     2     0     0    22     0\n",
       "PRAD     0     0     0     0    33"
      ]
     },
     "execution_count": 78,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "cm_df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "122f5339-6de4-48e5-a8b6-dd821b2bac7c",
   "metadata": {},
   "outputs": [],
   "source": [
    "#visualize via seaborn"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 74,
   "id": "e6d81bd9-26a2-450a-948c-b61118c92955",
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(50.722222222222214, 0.5, 'True Labels')"
      ]
     },
     "execution_count": 74,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "sns.heatmap(cm_df,annot=True,cmap='Blues')\n",
    "plt.xlabel('Predicted Labels')\n",
    "plt.ylabel('True Labels')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "ec539f23-c896-4b14-9992-b8c86a2cf46e",
   "metadata": {},
   "outputs": [],
   "source": [
    "#visualize matrix directly"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 86,
   "id": "6296161c-e46b-4a19-a91d-21e076fee267",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "from sklearn.metrics import ConfusionMatrixDisplay\n",
    "\n",
    "disp = ConfusionMatrixDisplay.from_estimator(\n",
    "    estimator=RF,  # Replace with your trained model\n",
    "    X=X_test_selected,  # Test data\n",
    "    y=y_test,  # True labels\n",
    "    display_labels=labels,\n",
    "    cmap='Blues',\n",
    "    xticks_rotation='vertical'\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 96,
   "id": "20a71a9f-2c9d-4a0b-9a56-3f900c73da37",
   "metadata": {},
   "outputs": [],
   "source": [
    "#now, let's play around with ROC curves and generate 1 for each class\n",
    "#We will have to binarize the y_test labels because the probabilities(pred_prob) are calculated for each each class and need to match the formatting"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 102,
   "id": "039f371c-c24f-4956-ae44-ba6b26712e71",
   "metadata": {},
   "outputs": [],
   "source": [
    "y_test_binarized=label_binarize(y_test,classes=classes)    "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 106,
   "id": "80958531-be3c-430a-9c34-1b71fc620ff5",
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.metrics import roc_curve, auc"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 110,
   "id": "26f00dc3-6f16-4d57-83db-509a4e837afb",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fpr = {}\n",
    "tpr = {}\n",
    "thresh ={}\n",
    "roc_auc = dict()\n",
    "\n",
    "n_class = classes.shape[0]\n",
    "\n",
    "for i in range(n_class):    \n",
    "    fpr[i], tpr[i], thresh[i] = roc_curve(y_test_binarized[:,i], pred_prob[:,i])\n",
    "    roc_auc[i] = auc(fpr[i], tpr[i])\n",
    "\n",
    "    plt.plot(fpr[i], tpr[i], linestyle='--', \n",
    "             label='%s vs Rest (AUC=%0.2f)'%(labels[i],roc_auc[i]))\n",
    "\n",
    "plt.plot([0,1],[0,1],'b--')\n",
    "plt.xlim([0,1])\n",
    "plt.ylim([0,1.05])\n",
    "plt.title('Multiclass ROC curve')\n",
    "plt.xlabel('False Positive Rate')\n",
    "plt.ylabel('True Positive rate')\n",
    "plt.legend(loc='lower right')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "7e9b1808-4849-4909-acc0-d94b3bf99195",
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.12.4"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}