01-09-2016-rrfd-scikit-learn-pycon-2014-2.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "%matplotlib inline\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Lets do a imple machine learning problem.\n",
    "\n",
    "ML is about building programs with tunable parameters (typically an array of floating point values)that are adjusted automatically so as to improve their behavior by adapting to previously seen data.\n",
    "\n",
    "Most ML applications it is [n_samples x n_features], and are either supervised or unsupervised."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "from sklearn.linear_model import LinearRegression"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Estimator parameters: All the parameters of an estimator can be set when it is instantiated.\n",
    "\n",
    "Lets start off with LinearRegression"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "True\n"
     ]
    }
   ],
   "source": [
    "model = LinearRegression(normalize=True)\n",
    "print model.normalize"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "LinearRegression(copy_X=True, fit_intercept=True, n_jobs=1, normalize=True)\n"
     ]
    }
   ],
   "source": [
    "print model"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Now let's look at estimated parameters are estimated from the data at hand."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "x = np.array([0, 1, 2])\n",
    "y = np.array([0, 1, 2])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[<matplotlib.lines.Line2D at 0x107b81f90>]"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXkAAAEACAYAAABWLgY0AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nCczE0wAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x107b812d0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(x, y, marker = '.')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[[0]\n",
      " [1]\n",
      " [2]] \n",
      "[[0]\n",
      " [2]\n",
      " [4]]\n"
     ]
    }
   ],
   "source": [
    "# We're going to transpose the axis\n",
    "X = x[:, np.newaxis]\n",
    "x2 = np.array([0, 2, 4])\n",
    "X2 = x2[:, np.newaxis]\n",
    "print X, \"\\n\", X2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([ 0.5])"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model.fit(X2, y) \n",
    "model.coef_"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Remember supervised learning can be split into two categories: regression and classification."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "['target_names', 'data', 'target', 'DESCR', 'feature_names'] \n",
      "(150, 4) (150,) ['setosa' 'versicolor' 'virginica']\n"
     ]
    }
   ],
   "source": [
    "from sklearn import neighbors, datasets\n",
    "\n",
    "iris = datasets.load_iris()\n",
    "\n",
    "# This is the split between data and classifiers\n",
    "# The data and target are in floats, not actual strings.\n",
    "# This is how supervised learning happens.\n",
    "X, y = iris.data, iris.target\n",
    "\n",
    "print iris.keys(), '\\n', X.shape, y.shape, iris.target_names"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "['virginica']\n"
     ]
    }
   ],
   "source": [
    "# KNeighborsClassifier() --> Classifier implementing the k-nearest neighbors vote.\n",
    "# It will plot the decision boundaries for each class.\n",
    "knn = neighbors.KNeighborsClassifier(n_neighbors = 1)\n",
    "knn.fit(X, y)\n",
    "\n",
    "# Now are can prdict certain classifers\n",
    "# What kind of iris has 3cm x 5cm speal and 4cm x 2cm petals\n",
    "print iris.target_names[knn.predict([[3, 5, 4, 2]])]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "['virginica']\n"
     ]
    }
   ],
   "source": [
    "print iris.target_names[knn.predict([[1, 2, 5, 0]])]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "# But what do the actual decision boundaries look like?\n",
    "# This is a visualization of the knn, but just an image of it\n",
    "# not the acual code.\n",
    "\n",
    "# from fig_code import plot_iris_classification\n",
    "# plot_iris_classification(neighbors.KNeighborsClassifier, n_neighbors=3)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "['virginica'] ['setosa']\n"
     ]
    }
   ],
   "source": [
    "from sklearn.svm import SVC\n",
    "unknown_iris = [[3, 5, 4, 2]]\n",
    "unknown_iris2 = [[1, 6, 2, 5]]\n",
    "\n",
    "# Use SVC to train and predict the species of the unknown iris...\n",
    "\n",
    "# It could be any number really, but this is just the one that is used.\n",
    "\n",
    "print iris.target_names[knn.predict(unknown_iris)], iris.target_names[knn.predict(unknown_iris2)]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Lets try a regression model."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# Create some simple data\n",
    "np.random.seed(12)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[[ 0.08427267]\n",
      " [ 0.7325207 ]\n",
      " [ 0.63619999]\n",
      " [ 0.02790779]\n",
      " [ 0.30017006]\n",
      " [ 0.22085252]\n",
      " [ 0.05501999]\n",
      " [ 0.52324607]\n",
      " [ 0.41636966]\n",
      " [ 0.04821875]\n",
      " [ 0.57240452]\n",
      " [ 0.80351665]\n",
      " [ 0.11360172]\n",
      " [ 0.27711721]\n",
      " [ 0.64167809]\n",
      " [ 0.49273731]\n",
      " [ 0.50655608]\n",
      " [ 0.46144026]\n",
      " [ 0.89474307]\n",
      " [ 0.60577769]] [ 1.10460529  6.31190634  4.90954274  2.03230837  3.06029788  1.94629398\n",
      "  2.21558281  3.4264008   4.19268436  2.50230047  3.63376434  5.08835605\n",
      "  2.89686553  3.0540711   2.3960488   4.50742309  2.35340947  2.37475913\n",
      "  4.57896123  4.32935524]\n"
     ]
    }
   ],
   "source": [
    "# 20 rows and 1 col\n",
    "X = np.random.random(size=(20,1))\n",
    "\n",
    "# Squeeze() --> Remove single-dimensional entries from the \n",
    "# shape of an array.\n",
    "\n",
    "# 20 col and 1 row\n",
    "y = 3 * X.squeeze() + 2 + np.random.normal(size = 20)\n",
    "\n",
    "print X, y"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "LinearRegression(copy_X=True, fit_intercept=True, n_jobs=1, normalize=False) \n",
      "Model coefficient: 3.67803, and intercept: 1.79834\n"
     ]
    }
   ],
   "source": [
    "# Fit a linear regression to the data\n",
    "from sklearn.linear_model import LinearRegression\n",
    "\n",
    "model = LinearRegression(fit_intercept=True)\n",
    "# First you fit\n",
    "model.fit(X, y)\n",
    "\n",
    "print model, '\\n', \"Model coefficient: %.5f, and intercept: %.5f\" % (model.coef_, model.intercept_)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.collections.PathCollection at 0x108422b10>"
      ]
     },
     "execution_count": 50,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXEAAAEACAYAAABF+UbAAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nqWD/BxKkNVbTyG4gAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1082764d0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Plot the data and the model prediction\n",
    "\n",
    "# Create random numbers and transpose them.\n",
    "X_test = np.linspace(0, 1, 100)[:, np.newaxis]\n",
    "\n",
    "# Then you predict, based on the \"model\" you created in the previous cell\n",
    "y_test = model.predict(X_test)\n",
    "\n",
    "import pylab as plt\n",
    "\n",
    "plt.scatter(X, y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAW8AAAEACAYAAAB8nvebAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nAkjlLSISQCpvEZEAUnmLiATQ/wO3OGU4d82LrwAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x108c4e210>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(X.squeeze(), y, '.')\n",
    "\n",
    "# This is the regression line\n",
    "plt.plot(X_test.squeeze(), y_test);"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Lets take a look at unsupervised learning.\n",
    "\n",
    "Here the data has not labels and we are interested in finding similarities between the objects in question.\n",
    "\n",
    "This is a way to discovering labels from the data.\n",
    "\n",
    "UL comprises tasks such as dimensionality reduction, clustering, and density estimation. We can even use UL to determine combinations of the measurements which best display the structure of the data.\n",
    "\n",
    "You can use UL discoveries to then run this on SL methods."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 69,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Reduced dataset shape: (150, 2)\n"
     ]
    }
   ],
   "source": [
    "X, y = iris.data, iris.target\n",
    "\n",
    "# This will form the basis for the UL\n",
    "from sklearn.decomposition import PCA\n",
    "\n",
    "# Where going to extract n combinations of variables to visualize this.\n",
    "# We're only looking at X, and not the actual features\n",
    "num_components = 2\n",
    "pca = PCA(n_components = components)\n",
    "pca.fit(X)\n",
    "X_reduced = pca.transform(X)\n",
    "print \"Reduced dataset shape:\", X_reduced.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 66,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXwAAAEACAYAAACwB81wAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n+H9PebVFTf1W3wAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x108d86590>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.scatter(X_reduced[:, 0], X_reduced[:, 1], c=y);"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 71,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[ 0.36158968, -0.08226889,  0.85657211,  0.35884393],\n",
       "       [-0.65653988, -0.72971237,  0.1757674 ,  0.07470647]])"
      ]
     },
     "execution_count": 71,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pca.components_"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 70,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Meaning of the 2 components:\n",
      "0.36 x sepal length (cm) + -0.08 x sepal width (cm) + 0.86 x petal length (cm) + 0.36 x petal width (cm)\n",
      "-0.66 x sepal length (cm) + -0.73 x sepal width (cm) + 0.18 x petal length (cm) + 0.07 x petal width (cm)\n"
     ]
    }
   ],
   "source": [
    "# But what does that mean?\n",
    "\n",
    "print \"Meaning of the %d components:\" % num_components\n",
    "for component in pca.components_:\n",
    "    print \" + \".join(\"%.2f x %s\" % (value, name)\n",
    "                     for value, name in zip(component, iris.feature_names))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Now we should look at clustering"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 2",
   "language": "python",
   "name": "python2"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 2
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython2",
   "version": "2.7.11"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 0
}