L1c.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Liniowy model"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Tworzymy nasz model liniowy odróżniający gatunek irysów Setosa od innych!"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# Standardowe importowanie pakietów\n",
    "import numpy as np\n",
    "import matplotlib.pylab as plt\n",
    "%matplotlib inline\n",
    "from sklearn.datasets import load_iris"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# Załadujmy jeszcze raz dane iris\n",
    "iris = load_iris()\n",
    "X,Y = iris.data[:,0:2], iris.target\n",
    "Y = Y==0 # Troche magii żeby przewidywał tylko Iris Setosa albo \"nie Iris Setosa\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array(['setosa', 'versicolor', 'virginica'], \n",
       "      dtype='|S10')"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "iris.target_names # Setosa jest pierwsza (indeks 0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "LinearSVC(C=1.0, class_weight=None, dual=True, fit_intercept=True,\n",
       "     intercept_scaling=1, loss='squared_hinge', max_iter=1000,\n",
       "     multi_class='ovr', penalty='l2', random_state=None, tol=0.0001,\n",
       "     verbose=0)"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Ok! Dopasowywujemy model. \n",
    "from sklearn.svm import LinearSVC\n",
    "# Regresja liniowa. Jeden z wielu modeli, których końcowym wynikiem jest linia prosta :)\n",
    "pierwszy_model = LinearSVC() \n",
    "pierwszy_model.fit(X, Y) # Wspominany interfejs! Po zrobieniu fit mamy linię prostą"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[-1.54140519,  2.28320629]])"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pierwszy_model.coef_"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### b) Model machine learningowy (*pierwszy_model* w naszym kodzie) można streścić jako:\n",
    "\n",
    "\"Jeśli -1.541405 \\* **długość_płatka** + 2.28320578 \\* **szerokość_płatka** >= 0 to jest to kwiatek gatunku *Iris Setosa*\"\n",
    "\n",
    "Nie uwierzycie ile modeli jest tak prostych :)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/kudkudak/anaconda/lib/python2.7/site-packages/sklearn/utils/validation.py:386: DeprecationWarning: Passing 1d arrays as data is deprecated in 0.17 and willraise ValueError in 0.19. Reshape your data either using X.reshape(-1, 1) if your data has a single feature or X.reshape(1, -1) if it contains a single sample.\n",
      "  DeprecationWarning)\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "array([ True], dtype=bool)"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Można zapytać (predict w Estimator)\n",
    "pierwszy_model.predict([-0.109, 0.451]) # Kwiatek gatunku Iris Setosa! "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Ćwiczenie 6 (2 pkt)\n",
    "\n",
    "Zaprogramuj funkcje predict() ktora bedzie zwracac to co pierwszy_model.predict. Funkcja powinna przejsc testy i implementowac nastepujacy interfejs"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def predict(z):\n",
    "    # Tutaj twoj kod"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "for x in np.linspace(4.5, 8, 10):\n",
    "    for y in np.linspace(4, 10s, 10):\n",
    "        assert pierwszy_model.predict([[x, y]]) == predict([x, y]), \"Failed on {},{}\".format(x, y)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Wizualizacja nauczonego modelu"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.text.Text at 0x11154fd10>"
      ]
     },
     "execution_count": 28,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAWwAAAELCAYAAAD9brxbAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1108e5210>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Tworzymy siatkę punktów w odstępie 1e-2\n",
    "h = 1e-2\n",
    "x_min, x_max = X[:, 0].min() - 1, X[:, 0].max() + 1\n",
    "y_min, y_max = X[:, 1].min() - 1, X[:, 1].max() + 1\n",
    "xx, yy = np.meshgrid(np.arange(x_min, x_max, h),\n",
    "                     np.arange(y_min, y_max, h))\n",
    "\n",
    "# Przewidujemy\n",
    "Z = pierwszy_model.predict(np.c_[xx.ravel(), yy.ravel()])\n",
    "Z = Z.reshape(xx.shape)\n",
    "\n",
    "# Rysujemy ładny wykres \n",
    "plt.contourf(xx, yy, Z, cmap=plt.cm.Paired, alpha=0.8)\n",
    "plt.scatter(X[:, 0], X[:, 1], c=Y, cmap=plt.cm.Paired)\n",
    "plt.xlabel('Sepal length')\n",
    "plt.ylabel('Sepal width')\n",
    "plt.xlim(xx.min(), xx.max())\n",
    "plt.ylim(yy.min(), yy.max())\n",
    "plt.xticks(())\n",
    "plt.yticks(())\n",
    "plt.title(\"pierwszy_model\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Podstawy machine learningu"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Podsumujmy co wiemy:\n",
    "    \n",
    "* Machine Learning zajmuje się tworzeniem modeli, które uczymy na podstawie danych\n",
    "* Scikit-learn udostępnia wiele modeli, uczymy je za pomocą funkcji **fit**, przewidujemy za pomocą funkcji **predict**\n",
    "* Ważną klasą modeli są modele liniowe na których się skupimy\n",
    "* Znamy podstawy numpy i matplotlib\n",
    "\n",
    "Wprowadzimy jeszcze 2 pojęcia:\n",
    "* Testowanie modelu\n",
    "* Hiperparametry\n",
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### a) Testowanie modelu\n",
    "\n",
    "Chcemy wiedzieć jak dobry jest model który mamy. W tym celu sobie symulujemy, i dzielimy zbiór danych na ** zbiór trenujący ** i ** zbiór testujący **. Uczymy się na zbiorze trenującym, testujemy na testującym. Zwykle podsumowywujemy wynik modelu jedną liczbą, np. **dokładność** (ang. *accuracy*) czyli ilość procent poprawnie sklasyfikowanych przykładów\n",
    "\n",
    "Przetestujmy jak sobie radzi nasz LogisticRegression na zbiorze Iris. Teraz weźmiemy wszystkie przykłady do nauki"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# Załadujmy jeszcze raz dane iris\n",
    "iris = load_iris()\n",
    "X,Y = iris.data, iris.target"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# Dzielimy (scikit-learn)\n",
    "from sklearn.cross_validation import train_test_split\n",
    "X_train, X_test, Y_train, Y_test = train_test_split(X,Y, random_state=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Pierwszy przykład trenujący:  [ 6.5  2.8  4.6  1.5] Klasa:  1\n"
     ]
    }
   ],
   "source": [
    "print \"Pierwszy przykład trenujący: \", X_train[0], \"Klasa: \", Y_train[0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "drugi_model = LinearSVC(C=1) \n",
    "drugi_model.fit(X_train, Y_train) # Wspominany interfejs! Po zrobieniu fit mamy linię prostą\n",
    "Y_test_predicted = drugi_model.predict(X_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Przewidywania na pierwszych 10 przykladach:  [0 1 1 0 2 2 2 0 0 2]\n"
     ]
    }
   ],
   "source": [
    "print \"Przewidywania na pierwszych 10 przykladach: \", Y_test_predicted[0:10]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Dokładność modelu to:  89.4736842105 %\n"
     ]
    }
   ],
   "source": [
    "import sklearn\n",
    "print \"Dokładność modelu to: \",100*sklearn.metrics.accuracy_score(Y_test, Y_test_predicted), \"%\""
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### c) Hiperparametry\n",
    "\n",
    "Każdy model opisywany jest hiperparametrami. Przykładowo w regresji liniowej możemy podać \"fit_intercept\" co oznacza czy linia ma przechodzi przez środek współrzędnych czy nie. Innym parametrem jest C - intuicyjnie opisuje prostote modelu"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Cwiczenie 7 (2 pkt) \n",
    "\n",
    "1. Znalezc C ktore na X_test, Y_test ma 100%. \n",
    "2. Narysowac wykres, na osi X wartosc C, na osi Y dokladnosc dla danego C"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Dokładność modelu to:  84.2105263158 %\n"
     ]
    }
   ],
   "source": [
    "# Sprawdzanie C=0.1\n",
    "drugi_model = LinearSVC(C=0.1) \n",
    "drugi_model.fit(X_train, Y_train) # Wspominany interfejs! Po zrobieniu fit mamy linię prostą\n",
    "Y_test_predicted = drugi_model.predict(X_test)\n",
    "print \"Dokładność modelu to: \",100*sklearn.metrics.accuracy_score(Y_test, Y_test_predicted), \"%\""
   ]
  },
  {
   "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"
  },
  "toc": {
   "toc_cell": false,
   "toc_number_sections": true,
   "toc_threshold": 6,
   "toc_window_display": false
  }
 },
 "nbformat": 4,
 "nbformat_minor": 0
}