{"metadata":{"kernelspec":{"language":"python","display_name":"Python 3","name":"python3"},"language_info":{"pygments_lexer":"ipython3","nbconvert_exporter":"python","version":"3.6.4","file_extension":".py","codemirror_mode":{"name":"ipython","version":3},"name":"python","mimetype":"text/x-python"}},"nbformat_minor":4,"nbformat":4,"cells":[{"cell_type":"code","source":"##########################################################\n# 1. IMPORT ALL PACKAGES\n##########################################################\nimport pandas as pd\nimport numpy as np\nfrom sklearn import svm\nfrom sklearn.model_selection import train_test_split\nfrom sklearn.metrics import multilabel_confusion_matrix\nimport math\nimport seaborn as sns\nfrom statistics import mean ","metadata":{"_cell_guid":"b1076dfc-b9ad-4769-8c92-a6c4dae69d19","_uuid":"8f2839f25d086af736a60e9eeb907d3b93b6e0e5","execution":{"iopub.execute_input":"2021-11-14T21:05:22.368924Z","iopub.status.busy":"2021-11-14T21:05:22.367690Z","iopub.status.idle":"2021-11-14T21:05:23.513505Z","shell.execute_reply":"2021-11-14T21:05:23.512494Z","shell.execute_reply.started":"2021-11-14T21:00:20.445134Z"},"papermill":{"duration":1.156343,"end_time":"2021-11-14T21:05:23.513728","exception":false,"start_time":"2021-11-14T21:05:22.357385","status":"completed"},"tags":[]},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"##########################################################\n# 2. LOAD TESTING AND TRAINING DATASET\n##########################################################\ndf_train = pd.read_csv(\"../input/digit-recognizer/train.csv\",header=0)# header 0 means the first row is name of the coloumn \ndf_test = pd.read_csv(\"../input/digit-recognizer/test.csv\",header=0)# header 0 means the first row is name of the coloumn \ndf_sample_submission = pd.read_csv(\"../input/digit-recognizer/sample_submission.csv\",header=0)# header 0 means the first row is name of the coloumn \n\n# View sample data\ndf_train.head(10) ","metadata":{"execution":{"iopub.execute_input":"2021-11-14T21:05:23.532965Z","iopub.status.busy":"2021-11-14T21:05:23.532329Z","iopub.status.idle":"2021-11-14T21:05:28.356548Z","shell.execute_reply":"2021-11-14T21:05:28.357018Z","shell.execute_reply.started":"2021-11-14T21:00:21.651283Z"},"papermill":{"duration":4.836451,"end_time":"2021-11-14T21:05:28.357206","exception":false,"start_time":"2021-11-14T21:05:23.520755","status":"completed"},"tags":[]},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"##########################################################\n# 3. SHARE TO TEST AND TRAIN DATA\n##########################################################\n# We share train data because this \"sample_submission.csv\" only has one class/label\nx = df_train.iloc[:, 1:]\ny = df_train['label'].tolist()\n\n# # Select 10000 rows data as a testing dataset\nx_test = x.iloc[0:10000, :].values.astype('float32') # all pixel values \ny_test = y[0:10000] # Select label for testing data\nx_train = x.iloc[10000:, :].values.astype('float32') # all pixel values \ny_train = y[10000:]","metadata":{"execution":{"iopub.execute_input":"2021-11-14T21:05:28.380242Z","iopub.status.busy":"2021-11-14T21:05:28.379518Z","iopub.status.idle":"2021-11-14T21:05:28.451401Z","shell.execute_reply":"2021-11-14T21:05:28.451861Z","shell.execute_reply.started":"2021-11-14T21:00:26.407399Z"},"papermill":{"duration":0.087576,"end_time":"2021-11-14T21:05:28.452055","exception":false,"start_time":"2021-11-14T21:05:28.364479","status":"completed"},"tags":[]},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"##########################################################\n# 4. FEATURE NORMALIZATION FOR BOTH (TEST & TRAIN)\n##########################################################\n# Proceed to normalize the features because the pixel intensities are currently between the range of 0 and 255\nprint((min(x_train[2]), max(x_train[2])))\n\n# Normalizing the data\nx_train = x_train/255.0\nx_test = x_test/255.0\n\n# Printing the shape of the Datasets\nprint('x_train shape:', x_train.shape)\nprint(x_train.shape[0], 'train samples')\nprint(x_test.shape[0], 'test samples')","metadata":{"execution":{"iopub.execute_input":"2021-11-14T21:05:28.474006Z","iopub.status.busy":"2021-11-14T21:05:28.473354Z","iopub.status.idle":"2021-11-14T21:05:28.530453Z","shell.execute_reply":"2021-11-14T21:05:28.529942Z","shell.execute_reply.started":"2021-11-14T21:00:26.486710Z"},"papermill":{"duration":0.072148,"end_time":"2021-11-14T21:05:28.530595","exception":false,"start_time":"2021-11-14T21:05:28.458447","status":"completed"},"tags":[]},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"##########################################################\n# 5. TRAIN SVM ALGORITHM\n##########################################################\n\nmodel_clf = svm.SVC(C=10, kernel='rbf', gamma=0.01)\n\n# Train the svm algorithm\nmodel_clf.fit(x_train, y_train)","metadata":{"execution":{"iopub.execute_input":"2021-11-14T21:05:28.550036Z","iopub.status.busy":"2021-11-14T21:05:28.548936Z","iopub.status.idle":"2021-11-14T21:08:20.354179Z","shell.execute_reply":"2021-11-14T21:08:20.354639Z","shell.execute_reply.started":"2021-11-14T21:00:26.547689Z"},"papermill":{"duration":171.817641,"end_time":"2021-11-14T21:08:20.354811","exception":false,"start_time":"2021-11-14T21:05:28.537170","status":"completed"},"tags":[]},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"##########################################################\n# 6. APPLY THE TRAINED LEARNER TO TEST NEW DATA\n##########################################################\n# Apply the trained perceptron to make prediction of test data\ny_pred = model_clf.predict(x_test)","metadata":{"execution":{"iopub.execute_input":"2021-11-14T21:08:20.373026Z","iopub.status.busy":"2021-11-14T21:08:20.372393Z","iopub.status.idle":"2021-11-14T21:09:59.290071Z","shell.execute_reply":"2021-11-14T21:09:59.289467Z","shell.execute_reply.started":"2021-11-14T21:03:17.789691Z"},"papermill":{"duration":98.928764,"end_time":"2021-11-14T21:09:59.290262","exception":false,"start_time":"2021-11-14T21:08:20.361498","status":"completed"},"tags":[]},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"##########################################################\n# 7. MULTI-CLASS CONFUSION MATRIX FOR EACH CLASS\n##########################################################\n# Actual and predicted classes\nlst_actual_class = y_test\nlst_predicted_class = y_pred\n\n# Class = digit labels 0, 1, 2, 3, 4, 5, 6, 7, 8, 9\nlst_classes = [0, 1, 2, 3, 4, 5, 6, 7, 8, 9] # Must in order\n\n# Compute multi-class confusion matrix\narr_out_matrix = multilabel_confusion_matrix(lst_actual_class, lst_predicted_class, labels=lst_classes)\n\n# Temp store results\nstore_sens = [];\nstore_spec = [];\nstore_acc = [];\nstore_bal_acc = [];\nstore_prec = [];\nstore_fscore = [];\nstore_mcc = [];\n\n# Loop for each taget label\nfor no_class in range(len(lst_classes)):\n    arr_data = arr_out_matrix[no_class];\n    print(\"Predicted Performance of Digit Label/Class: {0}\".format(no_class));\n\n    tp = arr_data[1][1]\n    fp = arr_data[0][1]\n    tn = arr_data[0][0]\n    fn = arr_data[1][0]\n\n    sensitivity = round(tp/(tp+fn), 3);\n    specificity = round(tn/(tn+fp), 3);\n    accuracy = round((tp+tn)/(tp+fp+tn+fn), 3);\n    balanced_accuracy = round((sensitivity+specificity)/2, 3);\n    precision = round(tp/(tp + fp), 3)\n    fscore = round((2 * ((precision * sensitivity) / (precision + sensitivity))), 3)\n    mcc = round((((tp * tn)-(fp * fn))/ math.sqrt((tp + fp)*(tp + fn)*(tn + fp)*(tn + fn))),3)\n\n    store_sens.append(sensitivity);\n    store_spec.append(specificity);\n    store_acc.append(accuracy);\n    store_bal_acc.append(balanced_accuracy);\n    store_prec.append(precision);\n    store_fscore.append(fscore);\n    store_mcc.append(mcc);\n    \n    print(\"TP={0}, FP={1}, TN={2}, FN={3}\".format(tp, fp, tn, fn));\n    print(\"Sensitivity: {0}\".format(sensitivity));\n    print(\"Specificity: {0}\".format(specificity));\n    print(\"Accuracy: {0}\".format(accuracy));\n    print(\"Balanced Accuracy: {0}\".format(balanced_accuracy));\n    print(\"Precision: {0}\".format(precision));\n    print(\"F1-Score: {0}\".format(fscore))\n    print(\"MCC: {0}\\n\".format(mcc))","metadata":{"execution":{"iopub.execute_input":"2021-11-14T21:09:59.319388Z","iopub.status.busy":"2021-11-14T21:09:59.318619Z","iopub.status.idle":"2021-11-14T21:09:59.338818Z","shell.execute_reply":"2021-11-14T21:09:59.338322Z","shell.execute_reply.started":"2021-11-14T21:04:56.586387Z"},"papermill":{"duration":0.041456,"end_time":"2021-11-14T21:09:59.338996","exception":false,"start_time":"2021-11-14T21:09:59.297540","status":"completed"},"tags":[]},"execution_count":null,"outputs":[]}]}