Scratch_CNN_2.2.ipynb

{"nbformat":4,"nbformat_minor":0,"metadata":{"colab":{"name":"Scratch_CNN_2.2.ipynb","provenance":[{"file_id":"1cpTnPSjhTgqUk82bMzEyxW8OpC7Kk2xJ","timestamp":1610809446438}],"collapsed_sections":[],"toc_visible":true},"kernelspec":{"name":"python3","display_name":"Python 3"},"accelerator":"GPU"},"cells":[{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"caRzHsGdMGOc","executionInfo":{"status":"ok","timestamp":1611767517624,"user_tz":-60,"elapsed":48425,"user":{"displayName":"FABIANA RITORTI","photoUrl":"","userId":"12810380093707210794"}},"outputId":"d7b06b5b-ee74-4433-9c91-e7affec94249"},"source":["from google.colab import drive\n","drive.mount('/content/drive')"],"execution_count":null,"outputs":[{"output_type":"stream","text":["Mounted at /content/drive\n"],"name":"stdout"}]},{"cell_type":"code","metadata":{"id":"nsPWVZGrNXQz"},"source":["import os\n","import pandas as pd\n","import numpy as np\n","import tensorflow as tf\n","import random as rn\n","import matplotlib.pyplot as plt\n","from sklearn.metrics import accuracy_score\n","from tensorflow.keras.utils import to_categorical\n","\n","\n","\n","out_path = '/content/drive/My Drive/Ritorti[_Pasco]/project/numpy_data'\n","\n","BATCH_SIZE = 64\n","\n","\n","\n","\n"],"execution_count":null,"outputs":[]},{"cell_type":"markdown","metadata":{"id":"hxdL3fDCzwoA"},"source":["# **Dataset load**"]},{"cell_type":"code","metadata":{"id":"64znKI2FNv7U"},"source":["def load_training():\n","  images = np.load(os.path.join(out_path,'train_tensor.npy'))\n","  labels = np.load(os.path.join(out_path,'train_labels.npy'))\n","  return images,labels\n","\n","def load_testing():\n","  images = np.load(os.path.join(out_path,'public_test_tensor.npy'))\n","  labels = np.load(os.path.join(out_path,'public_test_labels.npy'))\n","  return images,labels\n","\n","\n"," \n","images_train, labels_train= load_training()\n","images_test, labels_test= load_testing()\n","\n","\n","\n","\n","\n","\n","\n","\n","\n","\n"],"execution_count":null,"outputs":[]},{"cell_type":"markdown","metadata":{"id":"MRwdSFhvz1GY"},"source":["# **Preprocessing**\n"]},{"cell_type":"markdown","metadata":{"id":"zNqLTu1tz2Os"},"source":["##Preparing training and test set for mass-calcification classification and normalization"]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"AHWmK7dNu2at","executionInfo":{"status":"ok","timestamp":1611767536687,"user_tz":-60,"elapsed":850,"user":{"displayName":"FABIANA RITORTI","photoUrl":"","userId":"12810380093707210794"}},"outputId":"4d185031-c805-4312-ed14-5a27dc046794"},"source":["#BENIGN - MALIGNANT\n","train_zipped = zip(images_train,labels_train)\n","test_zipped = zip(images_test,labels_test)\n","train_images_ben_mal,train_labels_ben_mal = [],[]\n","test_images_ben_mal,test_labels_ben_mal = [],[]\n","\n","#training\n","for im,la in train_zipped:\n","  if la == 1 or la == 3: #benign -> 0\n","    train_images_ben_mal.append(im)\n","    train_labels_ben_mal.append(0)\n","  elif la == 2 or la == 4: #malignant -> 1\n","    train_images_ben_mal.append(im)\n","    train_labels_ben_mal.append(1)\n","\n","#testing\n","for im,la in test_zipped:\n","  if la == 1 or la == 3: #benign -> 0\n","    test_images_ben_mal.append(im)\n","    test_labels_ben_mal.append(0)\n","  elif la == 2 or la == 4: #malignant -> 1\n","    test_images_ben_mal.append(im)\n","    test_labels_ben_mal.append(1)\n","\n","train_images_ben_mal = np.array(train_images_ben_mal)\n","train_labels_ben_mal = np.array(train_labels_ben_mal)\n","test_images_ben_mal = np.array(test_images_ben_mal)\n","test_labels_ben_mal = np.array(test_labels_ben_mal)\n","print(f\"Shape training: {train_images_ben_mal.shape}\")\n","print(f\"Shape testing: {test_images_ben_mal.shape}\")\n","print(f\"Shape label training: {train_labels_ben_mal.shape}\")\n","print(f\"Shape label testing: {test_labels_ben_mal.shape}\")\n","\n","#Normalization\n","\n","train_images_ben_mal = train_images_ben_mal.reshape((2676, 150,150,1))\n","test_images_ben_mal = test_images_ben_mal.reshape((336, 150,150,1))\n","\n","train_images_ben_mal  = train_images_ben_mal .astype('float32') / 65535\n","test_images_ben_mal= test_images_ben_mal.astype('float32') / 65535\n","\n"],"execution_count":null,"outputs":[{"output_type":"stream","text":["Shape training: (2676, 150, 150)\n","Shape testing: (336, 150, 150)\n","Shape label training: (2676,)\n","Shape label testing: (336,)\n"],"name":"stdout"}]},{"cell_type":"code","metadata":{"id":"iRXBPGgeQZ5Q"},"source":["# Setting of parameters for reproducible results\n","os.environ[\"PYTHONHASHSEED\"]=\"0\"\n","np.random.seed(31)\n","rn.seed(14)\n","tf.random.set_seed(1234)"],"execution_count":null,"outputs":[]},{"cell_type":"markdown","metadata":{"id":"OdsTArph0MKt"},"source":["## Some of our dataset view"]},{"cell_type":"markdown","metadata":{"id":"wLZPoBGE0OAL"},"source":["Distribution of classes"]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/","height":265},"id":"wBwX0X-70Sxw","executionInfo":{"status":"ok","timestamp":1611767545033,"user_tz":-60,"elapsed":1204,"user":{"displayName":"FABIANA RITORTI","photoUrl":"","userId":"12810380093707210794"}},"outputId":"286ec38c-4a4d-47ac-aadd-fa0bdbc1b638"},"source":["plt.hist(train_labels_ben_mal)\n","plt.show()"],"execution_count":null,"outputs":[{"output_type":"display_data","data":{"image/png":"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\n","text/plain":["<Figure size 432x288 with 1 Axes>"]},"metadata":{"tags":[],"needs_background":"light"}}]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/","height":283},"id":"6tB9_sAZ0Y_z","executionInfo":{"status":"ok","timestamp":1611767548096,"user_tz":-60,"elapsed":1128,"user":{"displayName":"FABIANA RITORTI","photoUrl":"","userId":"12810380093707210794"}},"outputId":"f9587537-87e4-43ae-c9f4-ed08ed2e15c3"},"source":["print(train_images_ben_mal[0].dtype) \n","plt.hist(train_images_ben_mal[0][0]) \n","plt.show()"],"execution_count":null,"outputs":[{"output_type":"stream","text":["float32\n"],"name":"stdout"},{"output_type":"display_data","data":{"image/png":"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\n","text/plain":["<Figure size 432x288 with 1 Axes>"]},"metadata":{"tags":[],"needs_background":"light"}}]},{"cell_type":"markdown","metadata":{"id":"vKbGbevV0elU"},"source":["# **Scratch network**"]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"RM3lT-e-Q_YS","executionInfo":{"status":"ok","timestamp":1611336711996,"user_tz":-60,"elapsed":828,"user":{"displayName":"FABIANA RITORTI","photoUrl":"","userId":"12810380093707210794"}},"outputId":"3af1cd0f-0785-408a-b26b-f62997e01dd2"},"source":["#SMALL CONVNET WITH CONV2D AND MAXPOOLING2D\n","from tensorflow.keras import layers\n","from tensorflow.keras import models\n","\n","\n","\n","model = models.Sequential()\n","model.add(layers.Conv2D(32, (3, 3), activation='relu', input_shape=(150, 150, 1)))\n","model.add(layers.MaxPooling2D((2, 2)))\n","model.add(layers.Conv2D(64, (3, 3), activation='relu'))\n","model.add(layers.MaxPooling2D((2, 2)))\n","model.add(layers.Conv2D(128, (3, 3), activation='relu'))\n","model.add(layers.MaxPooling2D((2, 2)))\n","model.add(layers.Conv2D(128, (3, 3), activation='relu'))\n","model.add(layers.MaxPooling2D((2, 2)))\n","model.add(layers.Flatten())\n","model.add(layers.Dense(512, activation='relu'))\n","model.add(layers.Dense(1, activation='sigmoid'))\n","model.summary()\n","\n","\n","\n"],"execution_count":null,"outputs":[{"output_type":"stream","text":["Model: \"sequential_7\"\n","_________________________________________________________________\n","Layer (type)                 Output Shape              Param #   \n","=================================================================\n","conv2d_28 (Conv2D)           (None, 148, 148, 32)      320       \n","_________________________________________________________________\n","max_pooling2d_28 (MaxPooling (None, 74, 74, 32)        0         \n","_________________________________________________________________\n","conv2d_29 (Conv2D)           (None, 72, 72, 64)        18496     \n","_________________________________________________________________\n","max_pooling2d_29 (MaxPooling (None, 36, 36, 64)        0         \n","_________________________________________________________________\n","conv2d_30 (Conv2D)           (None, 34, 34, 128)       73856     \n","_________________________________________________________________\n","max_pooling2d_30 (MaxPooling (None, 17, 17, 128)       0         \n","_________________________________________________________________\n","conv2d_31 (Conv2D)           (None, 15, 15, 128)       147584    \n","_________________________________________________________________\n","max_pooling2d_31 (MaxPooling (None, 7, 7, 128)         0         \n","_________________________________________________________________\n","flatten_7 (Flatten)          (None, 6272)              0         \n","_________________________________________________________________\n","dense_14 (Dense)             (None, 512)               3211776   \n","_________________________________________________________________\n","dense_15 (Dense)             (None, 1)                 513       \n","=================================================================\n","Total params: 3,452,545\n","Trainable params: 3,452,545\n","Non-trainable params: 0\n","_________________________________________________________________\n"],"name":"stdout"}]},{"cell_type":"code","metadata":{"id":"h5-sOPLHSZFK"},"source":["from tensorflow.keras import optimizers\n","\n","\n","model.compile(loss='binary_crossentropy',\n","              optimizer= 'adam',\n","              metrics=['accuracy'])"],"execution_count":null,"outputs":[]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"5XeFxvILcwSr","executionInfo":{"status":"ok","timestamp":1611336766138,"user_tz":-60,"elapsed":50242,"user":{"displayName":"FABIANA RITORTI","photoUrl":"","userId":"12810380093707210794"}},"outputId":"3ed0fd82-63cf-4f61-d809-93b9c2f6ad25"},"source":["history = model.fit(train_images_ben_mal, train_labels_ben_mal,validation_split=0.2, epochs=30, batch_size=64, shuffle = True)"],"execution_count":null,"outputs":[{"output_type":"stream","text":["Epoch 1/30\n","34/34 [==============================] - 2s 54ms/step - loss: 0.7178 - accuracy: 0.5566 - val_loss: 0.6734 - val_accuracy: 0.6250\n","Epoch 2/30\n","34/34 [==============================] - 2s 50ms/step - loss: 0.6788 - accuracy: 0.5793 - val_loss: 0.6563 - val_accuracy: 0.6250\n","Epoch 3/30\n","34/34 [==============================] - 2s 48ms/step - loss: 0.6602 - accuracy: 0.6114 - val_loss: 0.6475 - val_accuracy: 0.6306\n","Epoch 4/30\n","34/34 [==============================] - 2s 48ms/step - loss: 0.6565 - accuracy: 0.6092 - val_loss: 0.6586 - val_accuracy: 0.6157\n","Epoch 5/30\n","34/34 [==============================] - 2s 47ms/step - loss: 0.6685 - accuracy: 0.5994 - val_loss: 0.6654 - val_accuracy: 0.6250\n","Epoch 6/30\n","34/34 [==============================] - 2s 47ms/step - loss: 0.6426 - accuracy: 0.6378 - val_loss: 0.6538 - val_accuracy: 0.6269\n","Epoch 7/30\n","34/34 [==============================] - 2s 47ms/step - loss: 0.6746 - accuracy: 0.5815 - val_loss: 0.6529 - val_accuracy: 0.6381\n","Epoch 8/30\n","34/34 [==============================] - 2s 47ms/step - loss: 0.6394 - accuracy: 0.6372 - val_loss: 0.6451 - val_accuracy: 0.6343\n","Epoch 9/30\n","34/34 [==============================] - 2s 47ms/step - loss: 0.6239 - accuracy: 0.6563 - val_loss: 0.6345 - val_accuracy: 0.6493\n","Epoch 10/30\n","34/34 [==============================] - 2s 47ms/step - loss: 0.6261 - accuracy: 0.6559 - val_loss: 0.6410 - val_accuracy: 0.6474\n","Epoch 11/30\n","34/34 [==============================] - 2s 47ms/step - loss: 0.6145 - accuracy: 0.6669 - val_loss: 0.6502 - val_accuracy: 0.6418\n","Epoch 12/30\n","34/34 [==============================] - 2s 47ms/step - loss: 0.6446 - accuracy: 0.6339 - val_loss: 0.6351 - val_accuracy: 0.6698\n","Epoch 13/30\n","34/34 [==============================] - 2s 47ms/step - loss: 0.6290 - accuracy: 0.6453 - val_loss: 0.6439 - val_accuracy: 0.6604\n","Epoch 14/30\n","34/34 [==============================] - 2s 48ms/step - loss: 0.6389 - accuracy: 0.6366 - val_loss: 0.6358 - val_accuracy: 0.6511\n","Epoch 15/30\n","34/34 [==============================] - 2s 48ms/step - loss: 0.6204 - accuracy: 0.6563 - val_loss: 0.6635 - val_accuracy: 0.6213\n","Epoch 16/30\n","34/34 [==============================] - 2s 48ms/step - loss: 0.6239 - accuracy: 0.6612 - val_loss: 0.6424 - val_accuracy: 0.6511\n","Epoch 17/30\n","34/34 [==============================] - 2s 47ms/step - loss: 0.6297 - accuracy: 0.6435 - val_loss: 0.6365 - val_accuracy: 0.6437\n","Epoch 18/30\n","34/34 [==============================] - 2s 47ms/step - loss: 0.6298 - accuracy: 0.6430 - val_loss: 0.6338 - val_accuracy: 0.6530\n","Epoch 19/30\n","34/34 [==============================] - 2s 47ms/step - loss: 0.6221 - accuracy: 0.6613 - val_loss: 0.6449 - val_accuracy: 0.6511\n","Epoch 20/30\n","34/34 [==============================] - 2s 47ms/step - loss: 0.6233 - accuracy: 0.6642 - val_loss: 0.6446 - val_accuracy: 0.6437\n","Epoch 21/30\n","34/34 [==============================] - 2s 47ms/step - loss: 0.6100 - accuracy: 0.6786 - val_loss: 0.6448 - val_accuracy: 0.6567\n","Epoch 22/30\n","34/34 [==============================] - 2s 47ms/step - loss: 0.6070 - accuracy: 0.6729 - val_loss: 0.6467 - val_accuracy: 0.6623\n","Epoch 23/30\n","34/34 [==============================] - 2s 47ms/step - loss: 0.6086 - accuracy: 0.6612 - val_loss: 0.6560 - val_accuracy: 0.6642\n","Epoch 24/30\n","34/34 [==============================] - 2s 48ms/step - loss: 0.6067 - accuracy: 0.6596 - val_loss: 0.6553 - val_accuracy: 0.6623\n","Epoch 25/30\n","34/34 [==============================] - 2s 47ms/step - loss: 0.6062 - accuracy: 0.6709 - val_loss: 0.6583 - val_accuracy: 0.6511\n","Epoch 26/30\n","34/34 [==============================] - 2s 47ms/step - loss: 0.6097 - accuracy: 0.6660 - val_loss: 0.6528 - val_accuracy: 0.6586\n","Epoch 27/30\n","34/34 [==============================] - 2s 47ms/step - loss: 0.5955 - accuracy: 0.6601 - val_loss: 0.6748 - val_accuracy: 0.6399\n","Epoch 28/30\n","34/34 [==============================] - 2s 47ms/step - loss: 0.6112 - accuracy: 0.6574 - val_loss: 0.6714 - val_accuracy: 0.6511\n","Epoch 29/30\n","34/34 [==============================] - 2s 47ms/step - loss: 0.6010 - accuracy: 0.6653 - val_loss: 0.6594 - val_accuracy: 0.6455\n","Epoch 30/30\n","34/34 [==============================] - 2s 47ms/step - loss: 0.5999 - accuracy: 0.6821 - val_loss: 0.6818 - val_accuracy: 0.6511\n"],"name":"stdout"}]},{"cell_type":"markdown","metadata":{"id":"r65yG9kC0kHr"},"source":["Save the model"]},{"cell_type":"code","metadata":{"id":"B2PpIsJrlR8e"},"source":["#dir_name =\"/content/drive/My Drive/Ritorti[_Pasco]/models\"\n","#if not os.path.exists(dir_name):\n","#  os.makedirs(dir_name)\n","#model.save(os.path.join(dir_name,'benign_and_malign.h5'))"],"execution_count":null,"outputs":[]},{"cell_type":"markdown","metadata":{"id":"yF7MOP7v0ons"},"source":["# **Training and validation accuracy-loss diagrams**"]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/","height":545},"id":"M7R_79-bijlA","executionInfo":{"status":"ok","timestamp":1611336770108,"user_tz":-60,"elapsed":1078,"user":{"displayName":"FABIANA RITORTI","photoUrl":"","userId":"12810380093707210794"}},"outputId":"4b22c52f-1043-46bb-9f82-0c1cc8905835"},"source":["#PLOT THE LOSS AND ACCURACY\n","import matplotlib.pyplot as plt\n","\n","acc = history.history['accuracy']\n","val_acc = history.history['val_accuracy']\n","loss = history.history['loss']\n","val_loss = history.history['val_loss']\n","\n","epochs = range(len(acc))\n","\n","plt.plot(epochs, acc, 'r', label='Training accuracy')\n","plt.plot(epochs, val_acc, 'b', label='Validation accuracy')\n","plt.title('Training and validation accuracy')\n","plt.legend()\n","\n","plt.figure()\n","\n","plt.plot(epochs, loss, 'r', label='Training loss')\n","plt.plot(epochs, val_loss, 'b', label='Validation loss')\n","plt.title('Training and validation loss')\n","plt.legend()\n","\n","plt.show()"],"execution_count":null,"outputs":[{"output_type":"display_data","data":{"image/png":"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\n","text/plain":["<Figure size 432x288 with 1 Axes>"]},"metadata":{"tags":[],"needs_background":"light"}},{"output_type":"display_data","data":{"image/png":"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\n","text/plain":["<Figure size 432x288 with 1 Axes>"]},"metadata":{"tags":[],"needs_background":"light"}}]},{"cell_type":"markdown","metadata":{"id":"87Z9TN270vP3"},"source":["# **Evaluating the model with the test set**"]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"-eLgXerWiR5L","executionInfo":{"status":"ok","timestamp":1611336814368,"user_tz":-60,"elapsed":816,"user":{"displayName":"FABIANA RITORTI","photoUrl":"","userId":"12810380093707210794"}},"outputId":"9d8808e9-1519-4257-c779-5257b23590a5"},"source":["test_loss, test_acc = model.evaluate(test_images_ben_mal, test_labels_ben_mal)\n","print('test_acc:', test_acc)"],"execution_count":null,"outputs":[{"output_type":"stream","text":["11/11 [==============================] - 0s 10ms/step - loss: 0.6346 - accuracy: 0.6458\n","test_acc: 0.6458333134651184\n"],"name":"stdout"}]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"ze3sqzShikHy","executionInfo":{"status":"ok","timestamp":1611336937427,"user_tz":-60,"elapsed":827,"user":{"displayName":"FABIANA RITORTI","photoUrl":"","userId":"12810380093707210794"}},"outputId":"b69fa6ad-05d5-4922-fed7-c26be34b6372"},"source":["predictions = model.predict(test_images_ben_mal) # get the output for each sample\n","\n","predictions = np.round(predictions)\n","predictions = predictions.astype(int)\n","\n","unique, counts = np.unique(test_labels_ben_mal, return_counts=True)\n","print('true',dict(zip(unique, counts)))\n","unique, counts = np.unique(predictions, return_counts=True)\n","print('pred',dict(zip(unique, counts)))"],"execution_count":null,"outputs":[{"output_type":"stream","text":["true {0: 219, 1: 117}\n","pred {0: 218, 1: 118}\n"],"name":"stdout"}]},{"cell_type":"markdown","metadata":{"id":"HsFrTzv10078"},"source":["**We save the predictions in a txt file to use them in the ensable method**"]},{"cell_type":"code","metadata":{"id":"FmmZciIh4e8Z"},"source":["#import csv\n","#dir_data_store =\"/content/drive/My Drive/Ritorti[_Pasco]/data_store\"\n","#if not os.path.exists(dir_data_store):\n","#  os.makedirs(dir_data_store)\n","\n","#np.savetxt(os.path.join(dir_data_store,\"predictions_22.csv\"), predictions, delimiter=\",\")"],"execution_count":null,"outputs":[]},{"cell_type":"markdown","metadata":{"id":"cuDu-cFf04QH"},"source":["**Evaluation of the model with the following tools:**\n","\n","\n","1.   Confusion matrix\n","2.   Accuracy, precision, recall and f1-score\n","3.   ROC curve"]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"6vYcX8yWikeT","executionInfo":{"status":"ok","timestamp":1611336948397,"user_tz":-60,"elapsed":810,"user":{"displayName":"FABIANA RITORTI","photoUrl":"","userId":"12810380093707210794"}},"outputId":"37ac8bb8-9bab-47b0-b6a1-4559a40e8e60"},"source":["from sklearn import metrics\n","metrics.confusion_matrix(test_labels_ben_mal,predictions)"],"execution_count":null,"outputs":[{"output_type":"execute_result","data":{"text/plain":["array([[159,  60],\n","       [ 59,  58]])"]},"metadata":{"tags":[]},"execution_count":91}]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"itTITbSAimPw","executionInfo":{"status":"ok","timestamp":1611336951476,"user_tz":-60,"elapsed":1048,"user":{"displayName":"FABIANA RITORTI","photoUrl":"","userId":"12810380093707210794"}},"outputId":"94384098-38b1-4ab9-d262-137855842357"},"source":["# accuracy: (TP+TN)/(TP+FP+FN+TN): percentage of correctly predicted samples\n","metrics.accuracy_score(test_labels_ben_mal,predictions)"],"execution_count":null,"outputs":[{"output_type":"execute_result","data":{"text/plain":["0.6458333333333334"]},"metadata":{"tags":[]},"execution_count":92}]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"kEpOslWhin3T","executionInfo":{"status":"ok","timestamp":1611336952990,"user_tz":-60,"elapsed":879,"user":{"displayName":"FABIANA RITORTI","photoUrl":"","userId":"12810380093707210794"}},"outputId":"6b81fed5-7e7f-41f8-aa83-9e4a8bf7e667"},"source":["# precision: TP/(TP+FP): The precision is intuitively the ability of the classifier not to label as positive a sample that is negative.\n","metrics.precision_score(test_labels_ben_mal,predictions)"],"execution_count":null,"outputs":[{"output_type":"execute_result","data":{"text/plain":["0.4915254237288136"]},"metadata":{"tags":[]},"execution_count":93}]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"Q5SMjkw3iqIV","executionInfo":{"status":"ok","timestamp":1611336954141,"user_tz":-60,"elapsed":448,"user":{"displayName":"FABIANA RITORTI","photoUrl":"","userId":"12810380093707210794"}},"outputId":"7ee3bbf0-ab80-4e50-b28c-5576a63003b3"},"source":["# recall: TP/(TP+FN): The recall is intuitively the ability of the classifier to find all the positive samples.\n","metrics.recall_score(test_labels_ben_mal,predictions)"],"execution_count":null,"outputs":[{"output_type":"execute_result","data":{"text/plain":["0.49572649572649574"]},"metadata":{"tags":[]},"execution_count":94}]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"vyQj2ZkLiru_","executionInfo":{"status":"ok","timestamp":1611336955985,"user_tz":-60,"elapsed":492,"user":{"displayName":"FABIANA RITORTI","photoUrl":"","userId":"12810380093707210794"}},"outputId":"42d52b0b-dd48-4623-c545-1be1a561ef23"},"source":["# f1-score: 2*(precision*recall)/(precision + recall). The F1-score is the harmonic mean of precision and sensitivity\n","metrics.f1_score(test_labels_ben_mal,predictions)"],"execution_count":null,"outputs":[{"output_type":"execute_result","data":{"text/plain":["0.49361702127659574"]},"metadata":{"tags":[]},"execution_count":95}]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"g-LaY3MAitab","executionInfo":{"status":"ok","timestamp":1611336957621,"user_tz":-60,"elapsed":274,"user":{"displayName":"FABIANA RITORTI","photoUrl":"","userId":"12810380093707210794"}},"outputId":"e69988ed-d30e-42b5-f838-9e1e3de37e17"},"source":["# classification report: Build a text report showing the main classification metrics\n","print(metrics.classification_report(test_labels_ben_mal,predictions))"],"execution_count":null,"outputs":[{"output_type":"stream","text":["              precision    recall  f1-score   support\n","\n","           0       0.73      0.73      0.73       219\n","           1       0.49      0.50      0.49       117\n","\n","    accuracy                           0.65       336\n","   macro avg       0.61      0.61      0.61       336\n","weighted avg       0.65      0.65      0.65       336\n","\n"],"name":"stdout"}]},{"cell_type":"code","metadata":{"id":"L9qntkrhiwLx"},"source":["fpr,tpr,th = metrics.roc_curve(test_labels_ben_mal,predictions)\n","roc_auc = metrics.roc_auc_score(test_labels_ben_mal,predictions)"],"execution_count":null,"outputs":[]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/","height":295},"id":"ZX4A8b9MiwiK","executionInfo":{"status":"ok","timestamp":1611336962047,"user_tz":-60,"elapsed":873,"user":{"displayName":"FABIANA RITORTI","photoUrl":"","userId":"12810380093707210794"}},"outputId":"d326b050-2232-46f5-dc7c-76161de97bea"},"source":["from matplotlib import pyplot as plt\n","plt.figure()\n","plt.plot(fpr, tpr, label='ROC curve (area = %0.2f)' % roc_auc)\n","plt.plot([0, 1], [0, 1], color='navy', linestyle='--')\n","plt.xlim([0.0, 1.0])\n","plt.ylim([0.0, 1.0])\n","plt.xlabel('False Positive Rate')\n","plt.ylabel('True Positive Rate')\n","plt.title('Receiver operating characteristic example')\n","plt.legend(loc=\"lower right\")\n","#plt.savefig(os.path.join(dir_img,\"ROC_curve.png\"))\n","plt.show()"],"execution_count":null,"outputs":[{"output_type":"display_data","data":{"image/png":"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\n","text/plain":["<Figure size 432x288 with 1 Axes>"]},"metadata":{"tags":[],"needs_background":"light"}}]}]}