Project - Wardrobe Identification.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "from keras.datasets import fashion_mnist"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {},
   "outputs": [],
   "source": [
    "data = fashion_mnist.load_data()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(60000, 28, 28)"
      ]
     },
     "execution_count": 52,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data[0][0].shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(60000,)"
      ]
     },
     "execution_count": 53,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data[0][1].shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "metadata": {},
   "outputs": [],
   "source": [
    "trainX = data[0][0]\n",
    "trainY = data[0][1]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(60000, 28, 28)"
      ]
     },
     "execution_count": 55,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "trainX.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Target number mapping\n",
    "\n",
    "<img src=\"https://github.com/awantik/machine-learning-slides/blob/master/fashion-mnist.PNG?raw=true\" width=\"200px\">"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "metadata": {},
   "outputs": [],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "%matplotlib inline"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "3"
      ]
     },
     "execution_count": 59,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "trainY[3]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.image.AxesImage at 0x10df9095048>"
      ]
     },
     "execution_count": 58,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x10d802a8278>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.imshow(trainX[3],cmap='gray')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [],
   "source": [
    "X = trainX.reshape(60000,-1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(60000, 784)"
      ]
     },
     "execution_count": 28,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [],
   "source": [
    "Y = trainY"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(60000,)"
      ]
     },
     "execution_count": 30,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "Y.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.model_selection import train_test_split, GridSearchCV"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [],
   "source": [
    "trainX,testX, trainY, testY = train_test_split(X,Y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.decomposition import PCA"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.pipeline import make_pipeline\n",
    "from sklearn.svm import SVC"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {},
   "outputs": [],
   "source": [
    "pipeline = make_pipeline(\n",
    "                          PCA(n_components=32, whiten=True),\n",
    "                          SVC(C=1, gamma=100,kernel='rbf')\n",
    "                        )"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {},
   "outputs": [],
   "source": [
    "gs = GridSearchCV(pipeline,cv=5,param_grid={'svc__kernel':['poly','rbf'], 'svc__gamma':[10,100,1000]}, n_jobs=-1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "GridSearchCV(cv=5, error_score='raise-deprecating',\n",
       "       estimator=Pipeline(memory=None,\n",
       "     steps=[('pca', PCA(copy=True, iterated_power='auto', n_components=32, random_state=None,\n",
       "  svd_solver='auto', tol=0.0, whiten=True)), ('svc', SVC(C=1, cache_size=200, class_weight=None, coef0=0.0,\n",
       "  decision_function_shape='ovr', degree=3, gamma=100, kernel='rbf',\n",
       "  max_iter=-1, probability=False, random_state=None, shrinking=True,\n",
       "  tol=0.001, verbose=False))]),\n",
       "       fit_params=None, iid='warn', n_jobs=-1,\n",
       "       param_grid={'svc__kernel': ['poly', 'rbf'], 'svc__gamma': [10, 100, 1000]},\n",
       "       pre_dispatch='2*n_jobs', refit=True, return_train_score='warn',\n",
       "       scoring=None, verbose=0)"
      ]
     },
     "execution_count": 43,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "gs.fit(trainX, trainY)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'svc__gamma': 100, 'svc__kernel': 'poly'}"
      ]
     },
     "execution_count": 44,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "gs.best_params_"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.8477777777777777"
      ]
     },
     "execution_count": 45,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "gs.best_score_"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {},
   "outputs": [],
   "source": [
    "model = gs.best_estimator_"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 89,
   "metadata": {},
   "outputs": [],
   "source": [
    "pred = model.predict(testX)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 92,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.metrics import accuracy_score,confusion_matrix"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 91,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.8494666666666667"
      ]
     },
     "execution_count": 91,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "accuracy_score(y_pred=pred, y_true=testY)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 93,
   "metadata": {},
   "outputs": [],
   "source": [
    "import seaborn as sns; sns.set()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 101,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x10d8b0ae3c8>"
      ]
     },
     "execution_count": 101,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x10d8afe7860>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "sns.heatmap(confusion_matrix(y_pred=pred, y_true=testY), annot=True, fmt='d')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "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": "3.6.4"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}