LR_example.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# Boilerplate\n",
    "%matplotlib inline\n",
    "\n",
    "# Intel DAAL related imports\n",
    "from daal.data_management import HomogenNumericTable\n",
    "\n",
    "# Helpersfor getArrayFromNT and printNT. See utils.py\n",
    "from utils import *\n",
    "\n",
    "# Import numpy, matplotlib, seaborn\n",
    "import numpy as np\n",
    "import matplotlib\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "# Plotting configurations\n",
    "%config InlineBackend.figure_format = 'retina'\n",
    "plt.rcParams[\"figure.figsize\"] = (12, 9)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Linear Regression \n",
    "\n",
    "### Tutorial brief\n",
    "This tutorial is an example of using linear regression algorithms from pyDAAL to build predictive models. We use the well-studied Boston House Prices dataset to train a basic multiple linear regression model, and then a Ridge regression model. We  test the performance of these models in median house price prediction. The code for multiple linear regression model training and prediction is provided. You are required to write the code for Ridge regression model training and prediction.\n",
    "\n",
    "### Learning objectives\n",
    "* To understand and practice the typical code sequence of using pyDAAL for supervised learning.\n",
    "* To practice interactions and conversions between DAAL NumericTables and NumPy ndarrays.\n",
    "\n",
    "### Linear regression introduction\n",
    "Supervised learning involves training a model using data that has known responses, and then apply the model to predict responses for unseen data. In the case of **linear regression**, the model is linear. That is, \n",
    "\n",
    "$$ f_{\\beta}(X) = \\beta_0 + \\beta_1 x_1 + \\beta_2 x_2 + \\cdots + \\beta_k x_k $$\n",
    "\n",
    "$\\beta_0, \\beta_1, \\cdots, \\beta_k$ are the regression model coefficients. \n",
    "\n",
    "pyDAAL provides two linear regression algorithms:\n",
    "* **Multiple Linear Regression**: The model is trained by minimizing an objective function in the form of **Residual Sum of Squares**. pyDAAL supports two ways to train the model: 1) Normal Equation method, and 2) QR method.\n",
    "\n",
    "$$ \\sum \\limits_{i=1}^n\\left ( y_i - f_{\\beta}(X^i)\\right )^2 $$  \n",
    "* **Ridge Regression**: It is similar to multiple linear regression, but adds a regularization term to the objective function. The regularization term penalizes features with large values, thus makes the model less prone to overfitting. \n",
    "\n",
    "$$ \\sum \\limits_{i=1}^n\\left ( y_i - f_{\\beta}(X^i)\\right )^2 + \\alpha \\sum \\limits_{j=1}^k \\beta_j^2 $$\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### The Boston House Prices dataset\n",
    "The dataset has already been downloaded to the ./mldata folder. There are 506 rows and 14 columns. The first 13 columns are features (explanatory variables), and the last column is the dependent variable we try to make predictions for. Here's detailed information about this dataset, including descriptions of each feature:\n",
    "\n",
    "> Origin: \n",
    "\n",
    "> This dataset was taken from the StatLib library which is maintained at Carnegie Mellon University. \n",
    "\n",
    "> Creator: \n",
    "\n",
    "> Harrison, D. and Rubinfeld, D.L. \n",
    "> 'Hedonic prices and the demand for clean air', J. Environ. Economics & Management, vol.5, 81-102, 1978.\n",
    "\n",
    "> Data Set Information:\n",
    "\n",
    "> Concerns housing values in suburbs of Boston.\n",
    "\n",
    "\n",
    "> Attribute Information:\n",
    "\n",
    "> 1. CRIM: per capita crime rate by town \n",
    "> 2. ZN: proportion of residential land zoned for lots over 25,000 sq.ft. \n",
    "> 3. INDUS: proportion of non-retail business acres per town \n",
    "> 4. CHAS: Charles River dummy variable (= 1 if tract bounds river; 0 otherwise) \n",
    "> 5. NOX: nitric oxides concentration (parts per 10 million) \n",
    "> 6. RM: average number of rooms per dwelling \n",
    "> 7. AGE: proportion of owner-occupied units built prior to 1940 \n",
    "> 8. DIS: weighted distances to five Boston employment centres \n",
    "> 9. RAD: index of accessibility to radial highways \n",
    "> 10. TAX: full-value property-tax rate per \\$10,000 \n",
    "> 11. PTRATIO: pupil-teacher ratio by town \n",
    "> 12. B: 1000(Bk - 0.63)^2 where Bk is the proportion of blacks by town \n",
    "> 13. LSTAT: % lower status of the population \n",
    "> 14. MEDV: Median value of owner-occupied homes in \\$1000's\n",
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Feature selection\n",
    "To fit a good linear model, we usually need to carefully decide which features to be included as explanatory variables. Features that are not statistically significant should be excluded. Features that are highly correlated among themselves should also be excluded. Feature selection is an important topic by itself. But it is beyond the scope of this tutorial. For our purposes, let's just say we exclude `indus`, `age`, `zn`, `tax`, and `rad`. This leaves `crim`, `chas`, `nox`, `rm`, `dis`, `ptratio`, `b`, and `lstat` as the explanatory variables. These are corresponding to column indices 0, 3, 4, 5, 7, 10, 11, and 12 in the data. Later, when loading data into memory we'll choose only these columns."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Multiple linear regression code using pyDAAL\n",
    "The code below defines `class LinearRegression`, which encapsulates the basic multiple linear regression algorithm from pyDAAL. An instance of the class can be initialized using `normEq` or `qr`, corresponding to the two model training methods. Method `train` takes training data and the known response to learn a model and return it. Method `predict` takes a model object and test data, and make predictions. The code also defines function `getBetas` that returns the regression coefficients of a model; function `mse` that returns _mean squared errors_ (MSE) for given true values and fitted values; function `scores` that computes coefficient of determination _R-squared_."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "from daal.algorithms.linear_regression import training as lr_training\n",
    "from daal.algorithms.linear_regression import prediction as lr_prediction\n",
    "\n",
    "def getBetas(linear_model):\n",
    "    \"\"\"Return regression coefficients for a given linear model\n",
    "\n",
    "    Args:\n",
    "        linear_model: A trained model\n",
    "\n",
    "    Returns:\n",
    "        A n-by-(k+1) NumericTable contains betas, where n is the number of dependent\n",
    "        variables; k is the number of features (independent variables)\n",
    "    \"\"\"\n",
    "\n",
    "    return linear_model.getBeta()\n",
    "\n",
    "\n",
    "\n",
    "\n",
    "def mse(values, fitted_values):\n",
    "    \"\"\"Return Mean Squared Errors for fitted values w.r.t. true values\n",
    "\n",
    "    Args:\n",
    "        values: True values. NumericTable, nsamples-by-noutputs\n",
    "        fitted_values: True values. NumericTable, nsamples-by-noutputs\n",
    "\n",
    "    Returns:\n",
    "        A tuple contains MSE's\n",
    "    \"\"\"\n",
    "\n",
    "    y_t = getArrayFromNT(values)\n",
    "    y_p = getArrayFromNT(fitted_values)\n",
    "    rss = ((y_t - y_p) ** 2).sum(axis = 0)\n",
    "    mse = rss / y_t.shape[0]\n",
    "    return tuple(mse)\n",
    "\n",
    "\n",
    "\n",
    "\n",
    "def score(y_true, y_pred):\n",
    "    \"\"\"Compute R-squared and adjusted R-squared\n",
    "\n",
    "    Args:\n",
    "        y_true: True values. NumericTable, shape = (nsamples, noutputs)\n",
    "        y_pred: Predicted values. NumericTable, shape = (nsamples, noutputs)\n",
    "\n",
    "    Returns:\n",
    "        R2: A tuple with noutputs values\n",
    "    \"\"\"\n",
    "\n",
    "    y_t = getArrayFromNT(y_true)\n",
    "    y_p = getArrayFromNT(y_pred)\n",
    "    rss = ((y_t - y_p) ** 2).sum(axis = 0)\n",
    "    tss = ((y_t - y_t.mean(axis = 0)) ** 2).sum(axis = 0)\n",
    "    r2 = 1 - rss/tss\n",
    "    return tuple(r2)\n",
    "\n",
    "\n",
    "class LinearRegression:\n",
    "\n",
    "\n",
    "    def __init__(self, method = 'normEq'):\n",
    "        \"\"\"Initialize class parameters\n",
    "\n",
    "        Args:\n",
    "           method: The default method is based on Normal Equation ('normEq'). It\n",
    "           can also be QR method ('qr')\n",
    "        \"\"\"\n",
    "\n",
    "        if method != 'normEq' and method != 'qr':\n",
    "            warnings.warn(method + \n",
    "            ' method is not supported. Default method is used', \n",
    "            UserWarning)\n",
    "\n",
    "        self.method_ = method\n",
    "\n",
    "\n",
    "\n",
    "    def train(self, data, responses):\n",
    "        \"\"\"Train a Linear Regression model.\n",
    "\n",
    "        Args:\n",
    "            data: Training data\n",
    "            responses: Known responses to the training data\n",
    "\n",
    "        Returns:\n",
    "            A Linear Regression model object\n",
    "        \"\"\"\n",
    "\n",
    "        # Create a training algorithm object\n",
    "        if self.method_ == 'qr': \n",
    "            lr_training_alg = lr_training.Batch_Float64QrDense() \n",
    "        else:\n",
    "            lr_training_alg = lr_training.Batch_Float64NormEqDense() \n",
    "        # Set input\n",
    "        lr_training_alg.input.set(lr_training.data, data)\n",
    "        lr_training_alg.input.set(lr_training.dependentVariables, responses)\n",
    "        # Compute\n",
    "        results = lr_training_alg.compute()\n",
    "        # Return the trained model\n",
    "        return results.get(lr_training.model)\n",
    "\n",
    "\n",
    "\n",
    "    def predict(self, model, testdata, intercept = True):\n",
    "        \"\"\"Make prediction for unseen data using a trained model\n",
    "\n",
    "        Args:\n",
    "            model: A trained model\n",
    "            testdata: New data\n",
    "            intercept: A boolean to inidicate if intercept needs to be computed \n",
    "\n",
    "        Returns:\n",
    "            A NumericTable containing predicted responses \n",
    "        \"\"\"\n",
    "\n",
    "        # Create a prediction algorithm object\n",
    "        lr_prediction_alg = lr_prediction.Batch_Float64DefaultDense()\n",
    "        # Set input\n",
    "        lr_prediction_alg.input.setModel(lr_prediction.model, model)\n",
    "        lr_prediction_alg.input.setTable(lr_prediction.data, testdata)\n",
    "        # Set parameters\n",
    "        lr_prediction_alg.parameter.interceptFlag = intercept\n",
    "        # Compute\n",
    "        results = lr_prediction_alg.compute()\n",
    "        return results.get(lr_prediction.prediction)\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Linear regression model training and prediction for Boston houses prices\n",
    "We save the last 50 samples from the entire dataset as test data, and use the rest as training data. The code below uses the `LinearRegression` class defined above. After training a model, we use it to predict prices for the last 50 samples. The code also shows computation of R-squared using the `scores` function, and retrieve the regression coefficients (betas) from the model.\n",
    "\n",
    "The code below reads data from file `housing.data` and creates 4 NumericTables: training data (`boston_X_train`), training data ground truth (`boston_Y_train`), test data (`boston_X_test`), and test data ground truth (`boston_Y_test`). We use the `numpy.genfromtxt` function to read data from the file into ndarrays. Then, create NumericTables from the ndarrays. _Note that when creating NumericTables from ndarrays, it is critical the ndarrays must be C-contiguous in memory._ "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "collapsed": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(0.44116933893793719,)\n",
      " \n",
      " [[  2.87404604e+01  -5.00910527e-02   2.89388785e+00  -1.62558071e+01\n",
      "    4.39060434e+00  -1.18956541e+00  -9.05072623e-01   8.96346755e-03\n",
      "   -5.32682714e-01]]\n"
     ]
    }
   ],
   "source": [
    "# Use only `crim`, `chas`, `nox`, `rm`, `dis`, `ptratio`, `b`, and `lstat` for explanatory variables\n",
    "cols = (0,3,4,5,7,10,11,12)\n",
    "data = np.genfromtxt('./mldata/housing.data', dtype=np.double, usecols=cols)\n",
    "truth = np.genfromtxt('./mldata/housing.data', dtype=np.double, usecols=(13))\n",
    "\n",
    "# Training data and responses\n",
    "boston_X_train = HomogenNumericTable(data[:-50])\n",
    "boston_Y_train = HomogenNumericTable(truth[:-50].reshape(-1,1))\n",
    "\n",
    "# Test data and ground truth\n",
    "boston_X_test = HomogenNumericTable(data[-50:])\n",
    "boston_Y_test = HomogenNumericTable(truth[-50:].reshape(-1,1))\n",
    "\n",
    "# Use the LinearRegression class for model training and prediction\n",
    "regr = LinearRegression()\n",
    "model = regr.train(boston_X_train, boston_Y_train)\n",
    "predictions = regr.predict(model, boston_X_test)\n",
    "\n",
    "# Calculate the coefficient of determination\n",
    "R2 = score(boston_Y_test, predictions)\n",
    "print(R2)\n",
    "\n",
    "# Peek at the model (Betas)\n",
    "printNT(getBetas(model))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Plotting predicted values agains the ground truth\n",
    "To see if the model has done a good job, we plot the predicted values against the ground truth. If the model does a perfect job then all points on the plot should fall on a straight line. As we see, it's not quite the case. But still the predictions are close to true values in many cases."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "collapsed": true
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAABaMAAARKCAYAAABSP2ybAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x333641c240>"
      ]
     },
     "metadata": {
      "image/png": {
       "height": 549,
       "width": 721
      }
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "y = truth[-50:]\n",
    "predicted = getArrayFromNT(predictions)\n",
    "\n",
    "fig, ax = plt.subplots()\n",
    "ax.scatter(y,predicted)\n",
    "ax.plot([y.min(), y.max()], [y.min(), y.max()], 'k--', lw=4)\n",
    "ax.set_xlabel('Measured')\n",
    "ax.set_ylabel('Predicted')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Exercise: Implementing Ridge regression\n",
    "In this exercise, you are asked to implement a class for Ridge regression using pyDAAL. Ridge regression is basically linear regression with regularizations. The skeleton of the Ridge class code is given below. It has two methods, **train** and **predict**. The implementation for **train** is already provided. You only need to follow it as a recipe to implement the **predict** method."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "from daal.algorithms.ridge_regression import training as ridge_training\n",
    "from daal.algorithms.ridge_regression import prediction as ridge_prediction\n",
    "\n",
    "class Ridge:\n",
    "\n",
    "    def __init__(self):\n",
    "        pass\n",
    "\n",
    "\n",
    "\n",
    "    def train(self, data, responses, alpha = 1.0):\n",
    "        \"\"\"Train a Ridge Regression model.\n",
    "\n",
    "        Args:\n",
    "           data: Training data\n",
    "           responses: Known responses to the training data\n",
    "           alpha: Regularization parameter, a small positive value with default\n",
    "           1.0\n",
    "\n",
    "        Returns:\n",
    "            A Ridge Regression model object\n",
    "        \"\"\"\n",
    "\n",
    "        # Create a training algorithm object\n",
    "        ridge_training_alg = ridge_training.Batch_Float64DefaultDense() \n",
    "        # Set input\n",
    "        ridge_training_alg.input.set(ridge_training.data, data)\n",
    "        ridge_training_alg.input.set(ridge_training.dependentVariables, responses)\n",
    "        # Set parameter\n",
    "        alpha_nt = HomogenNumericTable(np.array([alpha], ndmin=2))\n",
    "        ridge_training_alg.parameter.ridgeParameters = alpha_nt\n",
    "        # Compute\n",
    "        results = ridge_training_alg.compute()\n",
    "        # Return the trained model\n",
    "        return results.get(ridge_training.model)\n",
    "\n",
    "\n",
    "\n",
    "    def predict(self, model, testdata, intercept = True):\n",
    "        \"\"\"Make prediction for unseen data using a trained model\n",
    "\n",
    "        Args:\n",
    "            model: A trained model\n",
    "            testdata: New data\n",
    "            intercept: A boolean to inidicate if intercept needs to be computed \n",
    "\n",
    "        Returns:\n",
    "            A NumericTable containing predicted responses \n",
    "        \"\"\"\n",
    "\n",
    "        # Create a prediction algorithm object\n",
    "        #\n",
    "        # YOUR CODE HERE\n",
    "        #\n",
    "        # The algorithm class you want to use is ridge_prediction.Batch_Float64DefaultDense\n",
    "        \n",
    "        \n",
    "        # Set input\n",
    "        #\n",
    "        # YOUR CODE HERE\n",
    "        #\n",
    "        # There are two pieces of input to be set: a pre-trained model and input data. You should\n",
    "        # use the 'input.setModelInput' and the 'input.setNumericTableInput' member methods of the\n",
    "        # algorithm object. The input IDs to use are 'ridge_prediction.model' and 'ridge_prediction.data'\n",
    "        # respectively.\n",
    "        \n",
    "    \n",
    "        # Set parameters\n",
    "        #\n",
    "        # YOUR CODE HERE\n",
    "        #\n",
    "        # There is only one parameter we need to set. 'parameter.interceptFlag', it is a member variable \n",
    "        # of the algorithm object. \n",
    "\n",
    "        \n",
    "        # Compute\n",
    "        #\n",
    "        # YOUR CODE HERE\n",
    "        #\n",
    "        # Call the 'compute' method of your algorithm object, and store the result to 'results'.\n",
    "        \n",
    "        return results.get(ridge_prediction.prediction)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Test your Ridge regression class using the same Boston House Prices dataset. Note that all data read in previous steps is still available. You can try different values of $\\alpha$ (regularization parameter) to see its effect."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "collapsed": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[ 0.38482985]\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAABaMAAARKCAYAAABSP2ybAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nSUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x3334164b38>"
      ]
     },
     "metadata": {
      "image/png": {
       "height": 549,
       "width": 721
      }
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "#from regression import *\n",
    "\n",
    "# Use the Ridge class for model training and prediction\n",
    "regr = Ridge()\n",
    "model = regr.train(boston_X_train, boston_Y_train, alpha = 1.0)\n",
    "predictions = regr.predict(model, boston_X_test)\n",
    "\n",
    "# Calculate the coefficient of determination\n",
    "R2 = score(boston_Y_test, predictions)\n",
    "print(R2)\n",
    "\n",
    "y = truth[-50:]\n",
    "predicted = getArrayFromNT(predictions)\n",
    "\n",
    "fig, ax = plt.subplots()\n",
    "ax.scatter(y,predicted)\n",
    "ax.plot([y.min(), y.max()], [y.min(), y.max()], 'k--', lw=4)\n",
    "ax.set_xlabel('Measured')\n",
    "ax.set_ylabel('Predicted')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Summary\n",
    "In this lab, we learned two widely used linear regression models: Multiple linear regression and Ridge regression. We saw how to apply them to the Boston House Prices dataset. We studied and practiced pyDAAL API for these two algorithms."
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "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": "3.5.2"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 0
}