nolearn.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# nolearn"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "from __future__      import division\n",
    "from IPython.display import display\n",
    "from matplotlib      import pyplot as plt\n",
    "%matplotlib inline\n",
    "\n",
    "import numpy  as np\n",
    "import pandas as pd\n",
    "import random, sys, os\n",
    "\n",
    "from sklearn.cross_validation import StratifiedKFold\n",
    "from sklearn.grid_search      import RandomizedSearchCV, GridSearchCV"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "SEED   = 97\n",
    "scale  = True \n",
    "minmax = False\n",
    "norm   = False\n",
    "nointercept = True\n",
    "engineering = False\n",
    "\n",
    "N_CLASSES = 2\n",
    "\n",
    "submission_filename = \"../submissions/submission_nolearn.csv\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "global early_stop_type\n",
    "early_stop_type = 'unset'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "ImportError: Conv2DLayer and MaxPool2DLayer used\n"
     ]
    }
   ],
   "source": [
    "import theano\n",
    "from theano import tensor as T\n",
    "\n",
    "import lasagne\n",
    "from lasagne                import layers, nonlinearities\n",
    "from lasagne.nonlinearities import rectify, softmax, sigmoid\n",
    "from lasagne.updates        import nesterov_momentum, adam, adagrad, rmsprop\n",
    "from lasagne.layers         import InputLayer, DenseLayer, DropoutLayer\n",
    "from lasagne.layers         import get_all_params\n",
    "\n",
    "try:\n",
    "    from lasagne.layers.cuda_covnet import Conv2DCCLayer    as Conv2DLayer\n",
    "    from lasagne.layers.cuda_covnet import MaxPool2DCCLayer as MaxPool2DLayer\n",
    "    from lasagne.layers             import Pool2DLayer\n",
    "    print(\"Conv2DCCLayer and MaxPool2DCCLayer used\")\n",
    "    DCC = True\n",
    "except ImportError:\n",
    "    print(\"ImportError: Conv2DLayer and MaxPool2DLayer used\")\n",
    "    DCC = False\n",
    "    Conv2DLayer    = lasagne.layers.Conv2DLayer\n",
    "    MaxPool2DLayer = lasagne.layers.MaxPool2DLayer\n",
    "    Pool2DLayer    = lasagne.layers.Pool2DLayer"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "from nolearn.lasagne           import NeuralNet, TrainSplit, BatchIterator\n",
    "from nolearn.lasagne           import objective\n",
    "from nolearn.lasagne           import PrintLayerInfo\n",
    "\n",
    "from nolearn.lasagne.visualize import plot_loss\n",
    "from nolearn.lasagne.visualize import plot_conv_weights\n",
    "from nolearn.lasagne.visualize import plot_conv_activity\n",
    "from nolearn.lasagne.visualize import plot_occlusion"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "cpu\n"
     ]
    }
   ],
   "source": [
    "print theano.config.device"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Load the training data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "from load_blood_data import load_blood_data\n",
    "\n",
    "y_train, X_train = load_blood_data(train=True, SEED   = SEED, \n",
    "                                               scale  = scale,\n",
    "                                               minmax = minmax,\n",
    "                                               norm   = norm,\n",
    "                                               nointercept = nointercept,\n",
    "                                               engineering = engineering)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Fit the model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "\"\"\"\n",
    "[Using convolutional neural nets to detect facial keypoints; Daniel Nouri]\n",
    "\"\"\"\n",
    "\n",
    "def float32(k):\n",
    "    \"\"\"\n",
    "    Cast a variable for inclusion in a theano.shared variable\n",
    "    for subsequent use by AdjustVariable\n",
    "    \"\"\"\n",
    "    return np.cast['float32'](k)\n",
    "\n",
    "class AdjustVariable(object):\n",
    "    \"\"\"\n",
    "    Change the 'name' parameter each epoch\n",
    "    on a linspace(start, stop) schedule\n",
    "    \"\"\"\n",
    "    def __init__(self, name, start=0.03, stop=0.001):\n",
    "        self.name = name\n",
    "        self.start, self.stop = start, stop\n",
    "        self.ls = None\n",
    "\n",
    "    def __call__(self, nn, train_history):\n",
    "        if self.ls is None:\n",
    "            self.ls = np.linspace(self.start, self.stop, nn.max_epochs)\n",
    "\n",
    "        epoch = train_history[-1]['epoch']\n",
    "        new_value = float32(self.ls[epoch - 1])\n",
    "        getattr(nn, self.name).set_value(new_value)\n",
    "        \n",
    "class EarlyStopping_loss(object):\n",
    "    \"\"\"\n",
    "    Stop the training process if the validation loss has not declined\n",
    "    for 'patience' epochs\n",
    "    The model returned has the weights from the epoch with the lowest validation loss.\n",
    "    \"\"\"\n",
    "    def __init__(self, patience=100):\n",
    "        self.patience         = patience\n",
    "        self.best_valid       = np.inf\n",
    "        self.best_valid_epoch = 0\n",
    "        self.best_weights     = None\n",
    "        global early_stop_type\n",
    "        early_stop_type = 'loss'\n",
    "        \n",
    "\n",
    "    def __call__(self, nn, train_history):\n",
    "        current_valid = train_history[-1]['valid_loss']\n",
    "        current_epoch = train_history[-1]['epoch']\n",
    "        if current_valid < self.best_valid:\n",
    "            self.best_valid = current_valid\n",
    "            self.best_valid_epoch = current_epoch\n",
    "            self.best_weights = nn.get_all_params_values()\n",
    "        elif self.best_valid_epoch + self.patience < current_epoch:\n",
    "            print(\"\\nEarly stopping.\")\n",
    "            print(\"Best valid loss was {:.6f} at epoch {}.\\n\".format(\n",
    "                self.best_valid, self.best_valid_epoch))\n",
    "            nn.load_params_from(self.best_weights)\n",
    "            raise StopIteration()\n",
    "            \n",
    "class EarlyStopping_acc(object):\n",
    "    \"\"\"\n",
    "    Stop the training process if the validation accuracy has not increased\n",
    "    for 'patience' epochs\n",
    "    The model returned has the weights from the epoch with the highest validation accuracy.\n",
    "    \"\"\"\n",
    "    def __init__(self, patience=100):\n",
    "        self.patience         = patience\n",
    "        self.best_valid_acc   = 0\n",
    "        self.best_valid_epoch = 0\n",
    "        self.best_weights     = None\n",
    "        global early_stop_type\n",
    "        early_stop_type = 'acc'\n",
    "        \n",
    "\n",
    "    def __call__(self, nn, train_history):\n",
    "        global early_stop\n",
    "        current_valid_acc = train_history[-1]['valid_accuracy']\n",
    "        current_epoch     = train_history[-1]['epoch']\n",
    "        if current_valid_acc > self.best_valid_acc:\n",
    "            self.best_valid_acc   = current_valid_acc\n",
    "            self.best_valid_epoch = current_epoch\n",
    "            self.best_weights     = nn.get_all_params_values()\n",
    "        elif self.best_valid_epoch + self.patience < current_epoch:\n",
    "            print(\"\\nEarly stopping.\")\n",
    "            print(\"Best valid accuracy was {:.6f} at epoch {}.\\n\".format(\n",
    "                self.best_valid_acc, self.best_valid_epoch))\n",
    "            nn.load_params_from(self.best_weights)\n",
    "            raise StopIteration()\n",
    "            \n",
    "def elastic_net(layers, lambdaL1=0., lambdaL2=0., *args, **kwargs):\n",
    "    losses = objective(layers, *args, **kwargs)\n",
    "    # get the layers' weights, but not the biases\n",
    "    weights = get_all_params(layers[-1], regularizable=True)\n",
    "    # sum of absolute weights for L1\n",
    "    sum_abs_weights = sum([abs(w).sum() for w in weights])\n",
    "    # sum of squared weights for L2\n",
    "    sum_squared_weights = sum([(w ** 2).sum() for w in weights])\n",
    "    # add weights to regular loss\n",
    "    losses += lambdaL1 * sum_abs_weights + lambdaL2 * sum_squared_weights\n",
    "    return losses\n",
    "\n",
    "class FlipBatchIterator(BatchIterator):\n",
    "\n",
    "    def transform(self, Xb, yb):\n",
    "        Xb, yb = super(FlipBatchIterator, self).transform(Xb, yb)\n",
    "\n",
    "        # Flip half of the images in this batch at random:\n",
    "        bs = Xb.shape[0]\n",
    "        indices = np.random.choice(bs, bs / 2, replace=False)\n",
    "        Xb[indices] = Xb[indices, :, :, ::-1]\n",
    "\n",
    "        return Xb, yb"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "\"\"\"\n",
    "[Understanding the difficulty of training deep feedforward neural networks; Xavier Glorot, Yoshua Bengio]\n",
    "\"\"\"\n",
    "\n",
    "def xavierReLU(fan_in,fan_out):\n",
    "    \"\"\"\n",
    "    Xavier Glorot and Yohua Bengio recommended weight initialization\n",
    "    for layers using a ReLU activation function\n",
    "    \n",
    "    W ~ U[ +/- sqrt(6 / (fan_in + fan_out)) ]\n",
    "    \"\"\"\n",
    "    import numpy as np\n",
    "    return np.sqrt(6. / (fan_in + fan_out))\n",
    "\n",
    "def xavierTanH(fan_in,fan_out):\n",
    "    \"\"\"\n",
    "    Xavier Glorot and Yohua Bengio recommended weight initialization\n",
    "    for layers using a tanh activation function\n",
    "    \n",
    "    W ~ U[ +/- 4 * sqrt(6 / (fan_in + fan_out)) ]\n",
    "    \"\"\"\n",
    "    import numpy as np\n",
    "    return 4. * np.sqrt(6. / (fan_in + fan_out))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {
    "collapsed": false,
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "CPU times: user 0 ns, sys: 0 ns, total: 0 ns\n",
      "Wall time: 295 µs\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "\"\"\"\n",
    "[Neural Networks and Deep Learning; Michael Nielsen ]\n",
    "\"\"\"\n",
    "\n",
    "random.seed(SEED)\n",
    "\n",
    "\n",
    "clf_layers=[\n",
    "            ('input',    InputLayer),\n",
    "    \n",
    "    \n",
    "#             ('dropout1', DropoutLayer),\n",
    "            ('dense1',   DenseLayer),\n",
    "    \n",
    "#             ('dropout2', DropoutLayer),\n",
    "#             ('dense2',   DenseLayer),\n",
    "    \n",
    "\n",
    "            ('output',   DenseLayer),\n",
    "            ]\n",
    "\n",
    "\n",
    "\n",
    "clf = NeuralNet(\n",
    "    \n",
    "    layers=clf_layers,\n",
    "    \n",
    "    regression=False,\n",
    "    \n",
    "    # --------------------------------------------------------------------------\n",
    "    \n",
    "    input_shape=(None, X_train.shape[1]), \n",
    "    \n",
    "    # Fully Connected\n",
    "    # ===============\n",
    "#     dropout1_p=0.20,\n",
    "    \n",
    "    dense1_num_units    = 500,\n",
    "    dense1_nonlinearity = lasagne.nonlinearities.LeakyRectify(leakiness=1./3.),\n",
    "    dense1_W = lasagne.init.Uniform(range = xavierReLU(X_train.shape[1], 500)),\n",
    "    dense1_b = lasagne.init.Constant(1.),\n",
    "    \n",
    "#     dropout2_p=0.02,\n",
    "    \n",
    "#     dense2_num_units    = 10,\n",
    "#     dense2_nonlinearity = lasagne.nonlinearities.LeakyRectify(leakiness=1./3.),\n",
    "#     dense2_W = lasagne.init.Uniform(range = xavierReLU(500, 10)),\n",
    "#     dense2_b = lasagne.init.Constant(1.),\n",
    "    \n",
    "\n",
    "    # Output\n",
    "    # ======\n",
    "    output_num_units    = N_CLASSES, \n",
    "    output_nonlinearity = nonlinearities.softmax,\n",
    "    output_W = lasagne.init.Normal(std=1./np.sqrt(500 + N_CLASSES), mean=0.0),\n",
    "    output_b = lasagne.init.Constant(1.),\n",
    "    \n",
    "    # --------------------------------------------------------------------------\n",
    "    \n",
    "    # Loss Function\n",
    "    # =============\n",
    "    objective_loss_function = lasagne.objectives.categorical_crossentropy,\n",
    "    \n",
    "    # L1 & L2 Regularization\n",
    "    # ======================\n",
    "    objective          = elastic_net,\n",
    "    objective_lambdaL1 = 0.0000,\n",
    "    objective_lambdaL2 = 0.0000,\n",
    "    \n",
    "    # Update Function\n",
    "    # ===============\n",
    "    update               = lasagne.updates.momentum,\n",
    "    update_learning_rate = theano.shared(float32(0.003)),\n",
    "    update_momentum      = theano.shared(float32(0.9)),\n",
    "    \n",
    "    # Early Stopping\n",
    "    # ==============\n",
    "    on_epoch_finished=[\n",
    "        AdjustVariable('update_learning_rate', start=0.003,  stop=0.0003),\n",
    "        AdjustVariable('update_momentum',      start=0.9,    stop=0.99),\n",
    "        \n",
    "        EarlyStopping_loss(patience = 50)\n",
    "        ],\n",
    "    \n",
    "    \n",
    "    # Mini-Batch Size\n",
    "    # ===============\n",
    "    batch_iterator_train = BatchIterator(batch_size=1), \n",
    "    batch_iterator_test  = BatchIterator(batch_size=1),\n",
    "\n",
    "    # Max Epochs\n",
    "    # ==========\n",
    "    max_epochs = 100,\n",
    "    \n",
    "    verbose=0,\n",
    "    train_split=TrainSplit(0.2, stratify=True)\n",
    "    )\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "# Neural Network with 4002 learnable parameters\n",
      "\n",
      "## Layer information\n",
      "\n",
      "  #  name      size\n",
      "---  ------  ------\n",
      "  0  input        5\n",
      "  1  dense1     500\n",
      "  2  output       2\n",
      "\n",
      "CPU times: user 404 ms, sys: 8 ms, total: 412 ms\n",
      "Wall time: 409 ms\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "\n",
    "clf.initialize()\n",
    "layer_info = PrintLayerInfo()\n",
    "layer_info(clf, None)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {
    "collapsed": false,
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Early stopping.\n",
      "Best valid loss was 0.549808 at epoch 25.\n",
      "\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "NeuralNet(X_tensor_type=None,\n",
       "     batch_iterator_test=<nolearn.lasagne.base.BatchIterator object at 0x7f3b006e4250>,\n",
       "     batch_iterator_train=<nolearn.lasagne.base.BatchIterator object at 0x7f3b006e4a50>,\n",
       "     custom_score=None,\n",
       "     dense1_W=<lasagne.init.Uniform object at 0x7f3b006e4450>,\n",
       "     dense1_b=<lasagne.init.Constant object at 0x7f3b006e4950>,\n",
       "     dense1_nonlinearity=<lasagne.nonlinearities.LeakyRectify object at 0x7f3b006e4610>,\n",
       "     dense1_num_units=500, input_shape=(None, 5),\n",
       "     layers=[('input', <class 'lasagne.layers.input.InputLayer'>), ('dense1', <class 'lasagne.layers.dense.DenseLayer'>), ('output', <class 'lasagne.layers.dense.DenseLayer'>)],\n",
       "     loss=None, max_epochs=100, more_params={},\n",
       "     objective=<function elastic_net at 0x7f3b1e848f50>,\n",
       "     objective_lambdaL1=0.0, objective_lambdaL2=0.0,\n",
       "     objective_loss_function=<function categorical_crossentropy at 0x7f3b08c6f488>,\n",
       "     on_epoch_finished=[<__main__.AdjustVariable object at 0x7f3b006e4550>, <__main__.AdjustVariable object at 0x7f3b006e4690>, <__main__.EarlyStopping_loss object at 0x7f3b006e44d0>],\n",
       "     on_training_finished=[], on_training_started=[],\n",
       "     output_W=<lasagne.init.Normal object at 0x7f3b006e4bd0>,\n",
       "     output_b=<lasagne.init.Constant object at 0x7f3b006e46d0>,\n",
       "     output_nonlinearity=<function softmax at 0x7f3b0c9af500>,\n",
       "     output_num_units=2, regression=False,\n",
       "     train_split=<nolearn.lasagne.base.TrainSplit object at 0x7f3b006e4790>,\n",
       "     update=<function momentum at 0x7f3b08c6faa0>,\n",
       "     update_learning_rate=<TensorType(float32, scalar)>,\n",
       "     update_momentum=<TensorType(float32, scalar)>,\n",
       "     use_label_encoder=False, verbose=0,\n",
       "     y_tensor_type=TensorType(int32, vector))"
      ]
     },
     "execution_count": 43,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# clf_params = {'verbose': 0, 'n_estimators': 10, 'criterion': 'gini'}\n",
    "# clf.set_params(**clf_params)\n",
    "clf.fit(X_train.values.astype(np.float32), y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# =================================\n",
    "# Create vectors of loss & accuracy\n",
    "# =================================\n",
    "train_loss = np.array([i[\"train_loss\"]     for i in clf.train_history_])\n",
    "valid_loss = np.array([i[\"valid_loss\"]     for i in clf.train_history_])\n",
    "valid_acc  = np.array([i[\"valid_accuracy\"] for i in clf.train_history_])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAwIAAAJsCAYAAABONRQmAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nvvVGbrdWXAAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f3afea195d0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, ax1 = plt.subplots(figsize=(12,10))\n",
    "\n",
    "# ===============================\n",
    "# Plot Validation & Training Loss\n",
    "# ===============================\n",
    "ax1.plot(valid_loss, linewidth=3, label=\"valid loss\", color='g', linestyle='-')\n",
    "ax1.plot(train_loss, linewidth=1, label=\"train loss\", color='b', linestyle='-')\n",
    "stop_type = ' no early stopping' \n",
    "if early_stop_type=='acc':  stop_type = ' early stopping for accuracy'\n",
    "if early_stop_type=='loss': stop_type = ' early stopping for loss'\n",
    "ax1.set_xlabel('Epochs, '+stop_type,fontsize=24)\n",
    "\n",
    "# Make the y-axis label and tick labels match the line color.\n",
    "ax1.set_ylabel('loss', color='b',fontsize=24)\n",
    "for tl in ax1.get_yticklabels():\n",
    "    tl.set_color('b')\n",
    "    \n",
    "#plt.yscale(\"log\")\n",
    "#-----------------\n",
    "\n",
    "ax1.axvline(np.argmin(valid_loss)+0, color='g', linestyle='--',label=\"Best Loss\")\n",
    "ax1.axvline(np.argmax(valid_acc)+0,  color='r', linestyle='-.',label=\"Best Accuracy\")\n",
    "ax1.legend(loc=\"upper center\")\n",
    "\n",
    "# ========================\n",
    "# Plot Validation Accuracy\n",
    "# ========================\n",
    "ax2 = ax1.twinx()  # share the x axis\n",
    "ax2.plot(valid_acc, linewidth=3, label=\"valid accuracy\", color='r',linestyle=':')\n",
    "ax2.set_ylabel('Validation Accuracy', color='r',fontsize=24)\n",
    "\n",
    "# Make the y-axis label and tick labels match the line color.\n",
    "for tl in ax2.get_yticklabels():\n",
    "    tl.set_color('r')\n",
    "ax2.legend(loc=\"lower center\")\n",
    "\n",
    "plt.grid()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Training set predictions"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "predicted_values = clf.predict(X_train.values.astype(np.float32))\n",
    "y_true, y_pred   = y_train, predicted_values\n",
    "\n",
    "predicted_probs  = clf.predict_proba(X_train.values.astype(np.float32))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[[409  29]\n",
      " [ 82  56]]\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAfsAAAHECAYAAAA+p4OqAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nouUS7CMiIlruB+6oQ+DHGcSrAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f3afeb74ed0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy 0.8073, mis-class rate 0.1927\n"
     ]
    }
   ],
   "source": [
    "from sklearn.metrics import confusion_matrix\n",
    "cm = confusion_matrix(y_true, y_pred, labels=None)\n",
    "print cm\n",
    "\n",
    "try:\n",
    "    from sklearn_utilities import plot_confusion_matrix\n",
    "except:\n",
    "    import imp, os\n",
    "    util = imp.load_source('sklearn_utilities', os.path.expanduser('~/Dropbox/Python/sklearn_utilities.py'))\n",
    "    from sklearn_utilities import plot_confusion_matrix\n",
    "\n",
    "\n",
    "plot_confusion_matrix(cm, ['Did not Donate','Donated'])\n",
    "\n",
    "accuracy = round(np.trace(cm)/float(np.sum(cm)),4)\n",
    "misclass = 1 - accuracy\n",
    "print(\"Accuracy {}, mis-class rate {}\".format(accuracy,misclass))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAmUAAAGJCAYAAADL4URDAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nwJnUtWwuTFwAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f3afeaf7dd0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "from sklearn.metrics import roc_curve\n",
    "from sklearn.metrics import roc_auc_score\n",
    "from sklearn.metrics import log_loss\n",
    "from sklearn.metrics import f1_score\n",
    "\n",
    "fpr, tpr, thresholds = roc_curve(y_true, y_pred, pos_label=None)\n",
    "\n",
    "\n",
    "plt.figure(figsize=(10,6))\n",
    "plt.plot([0, 1], [0, 1], 'k--')\n",
    "plt.plot(fpr, tpr)\n",
    "\n",
    "AUC = roc_auc_score(y_true, y_pred, average='macro')\n",
    "plt.text(x=0.6,y=0.4,s=\"AUC         {:.4f}\"\\\n",
    "         .format(AUC),\n",
    "        fontsize=16)\n",
    "\n",
    "plt.text(x=0.6,y=0.3,s=\"accuracy {:.2f}%\"\\\n",
    "         .format(accuracy*100),\n",
    "        fontsize=16)\n",
    "\n",
    "logloss = log_loss(y_true, y_pred)\n",
    "plt.text(x=0.6,y=0.2,s=\"LogLoss   {:.4f}\"\\\n",
    "         .format(logloss),\n",
    "        fontsize=16)\n",
    "\n",
    "f1 = f1_score(y_true, y_pred)\n",
    "plt.text(x=0.6,y=0.1,s=\"f1             {:.4f}\"\\\n",
    "         .format(f1),\n",
    "        fontsize=16)\n",
    "\n",
    "plt.xlabel('False positive rate')\n",
    "plt.ylabel('True positive rate')\n",
    "plt.title('ROC curve')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "nolearn                    ,        ,   0.8073 ,  6.6560 , 0.6698 , 0.5022\n"
     ]
    }
   ],
   "source": [
    "# format for scores.csv file\n",
    "import re\n",
    "algo = re.search(r\"submission_(.*?)\\.csv\", submission_filename).group(1)\n",
    "print(\"{: <26} ,        ,   {:.4f} ,  {:.4f} , {:.4f} , {:.4f} ,  , \"\\\n",
    "      .format(algo,accuracy,logloss,AUC,f1))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### confirm that *sklearn.metrics.log_loss* conforms to the contest formula"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "6.656"
      ]
     },
     "execution_count": 50,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "def log_loss(y_true, y_pred):\n",
    "    \"\"\"\n",
    "    logloss formula\n",
    "    \"\"\"\n",
    "    import scipy\n",
    "    import math\n",
    "    \n",
    "    assert len(y_true)==len(y_pred)\n",
    "    \n",
    "    n     = len(y_true)\n",
    "    comps = []\n",
    "    \n",
    "    # prevent log(0/1): y_pred \\in [epsilon, 0.999]\n",
    "    epsilon = 1e-15\n",
    "    y_pred = scipy.maximum(epsilon, y_pred)\n",
    "    y_pred = scipy.minimum(1-epsilon, y_pred)\n",
    "    \n",
    "    for i in range(n):\n",
    "        comp1 =    y_true[i]    * math.log(  y_pred[i],     math.e)\n",
    "        comp2 = (1 - y_true[i]) * math.log((1 - y_pred[i]), math.e)\n",
    "        comps.append(comp1+comp2)\n",
    "    \n",
    "    return round(-sum(comps)/float(n),4)\n",
    "\n",
    "log_loss(y_true,y_pred)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# --------------------------------------------------------------------------------------------"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "collapsed": true
   },
   "source": [
    "# Test Set Predictions"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Read the test data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "from load_blood_data import load_blood_data\n",
    "\n",
    "X_test, IDs = load_blood_data(train=False, SEED   = SEED, \n",
    "                                           scale  = scale,\n",
    "                                           minmax = minmax,\n",
    "                                           norm   = norm,\n",
    "                                           nointercept = nointercept,\n",
    "                                           engineering = engineering)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Predict the test set with the fitted model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "y_pred        = clf.predict(X_test.values.astype(np.float32))\n",
    "y_pred_probs  = clf.predict_proba(X_test.values.astype(np.float32))\n",
    "donate_probs  = [prob[1] for prob in y_pred_probs]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[1 0 0 0 1 1 0 0 0 0]\n",
      "[[ 0.48901987  0.51098013]\n",
      " [ 0.88643295  0.11356703]\n",
      " [ 0.69935817  0.30064186]\n",
      " [ 0.64642584  0.35357416]\n",
      " [ 0.44492528  0.55507475]\n",
      " [ 0.4067255   0.59327453]\n",
      " [ 0.64343876  0.35656124]\n",
      " [ 0.78362268  0.21637732]\n",
      " [ 0.99497014  0.00502988]\n",
      " [ 0.91420865  0.08579135]]\n",
      "[0.51098013, 0.11356703, 0.30064186, 0.35357416, 0.55507475, 0.59327453, 0.35656124, 0.21637732, 0.0050298837, 0.085791349]\n"
     ]
    }
   ],
   "source": [
    "print(y_pred[:10])\n",
    "print(y_pred_probs[:10])\n",
    "print(donate_probs[:10])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Create the submission file"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 78,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "assert len(IDs)==len(donate_probs)\n",
    "\n",
    "f = open(submission_filename, \"w\")\n",
    "\n",
    "f.write(\",Made Donation in March 2007\\n\")\n",
    "for ID, prob in zip(IDs, donate_probs):\n",
    "    f.write(\"{},{}\\n\".format(ID,prob))\n",
    "    \n",
    "f.close()"
   ]
  },
  {
   "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": "2.7.10"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 0
}