def load_base(path='', column_names=None, type=None): path = get_project_root() + '/mlfwk/datasets/' + path if type == 'csv': base_result = read_csv(path, names=column_names) return base_result if type == 'arff': base_result = arff.loadarff(path) base_result = DataFrame(base_result[0]) else: base_result = None return base_result
# utilizando o x_test e o y_test da ultima realização for c in [0, 1, 2]: plot_dict['classes'].update({ c: { 'X': x[where(y == c)[0]], 'point': point[c], 'marker': marker[c] } }) # #FFAAAA red # #AAAAFF blue coloring(plot_dict, ListedColormap(['#87CEFA', '#228B22', "#FF00FF"]), xlabel='SepalLengthCm', ylabel='SepalWidthCm', title='mapa de cores com Rede Perceptron - ACC: ' + str(metric_results['ACCURACY'].round(2)), xlim=[-0.1, 1.1], ylim=[-0.1, 1.1], path=get_project_root() + '/run/TR-04/IRIS/results/' + 'color_map_sepal_test.jpg', save=True) # print('dataset shape %s' % Counter(base[:, 2])) # ------------------ All points ------------------------------------------------------------------- x = array(base[:, :2]) y = array(out_of_c_to_label(base[:, 2:])) xx, yy = generate_space(x) space = c_[xx.ravel(), yy.ravel()] point = { 0: 'bo', 1: 'go', 2: 'mo'
results['realization'].append(realization) for type in ['ACCURACY', 'precision', 'recall', 'f1_score']: results[type].append(metric_results[type]) results['cf'].sort(key=lambda x: x[0], reverse=True) final_result['best_cf'].append(results['cf'][0][1]) best_acc_clf = results['cf'][0][2] best_acc = results['cf'][0][0] for type in ['ACCURACY', 'precision', 'recall', 'f1_score']: final_result[type].append(mean(results[type])) final_result['std ' + type].append(std(results[type])) print(DataFrame(final_result)) DataFrame(final_result).to_csv(get_project_root() + '/run/TR-00/ARTIFICIAL/results/' + 'result_knn.csv') for i in range(len(final_result['best_cf'])): plt.figure(figsize=(10, 7)) df_cm = DataFrame(final_result['best_cf'][i], index=[i for i in range(2)], columns=[i for i in range(2)]) sn.heatmap(df_cm, annot=True) plt.title('Matriz de connfusão do KNN com acurácia de ' + str(best_acc * 100) + "%") plt.xlabel('Valor Esperado') plt.ylabel('Valor Encontrado')
results['realization'].append(realization) for type in ['ACCURACY', 'precision', 'recall', 'f1_score']: results[type].append(metric_results[type]) for type in ['ACCURACY', 'precision', 'recall', 'f1_score']: final_result[type].append(mean(results[type])) final_result['std ' + type].append(std(results[type])) results['cf'].sort(key=lambda x: x[0], reverse=True) final_result['best_cf'].append(results['cf'][0][1]) best_acc_clf = results['cf'][0][2] best_acc = results['cf'][0][0] df_cm = DataFrame(results['cf'][0][1], index=[i for i in range(C)], columns=[i for i in range(C)]) sn.heatmap(df_cm, annot=True) plt.title('Matriz de connfusão do DMC com acurácia de ' + str(round(best_acc, 2) * 100) + "%") plt.xlabel('Valor Esperado') plt.ylabel('Valor Encontrado') path = get_project_root() + '/run/ML-00/COLUNA_3C/results/' plt.savefig(path + one_versus_others + "-conf_result_dmc.jpg") plt.show() print(DataFrame(final_result)) DataFrame(final_result).to_csv(get_project_root() + '/run/ML-00/COLUNA_3C/results/' + one_versus_others + '_result_dmc.csv')
final_result['alphas'].append(mean(results['alphas'])) for type in ['ACCURACY', 'precision', 'recall', 'f1_score']: final_result[type].append(mean(results[type])) final_result['std ' + type].append(std(results[type])) # ------------------------ PLOT ------------------------------------------------- for i in range(len(final_result['best_cf'])): plt.figure(figsize=(10, 7)) df_cm = DataFrame(final_result['best_cf'][i], index=[i for i in range(C)], columns=[i for i in range(C)]) sn.heatmap(df_cm, annot=True) plt.title('Matriz de connfusão dermatologia com raio de abertura: ' + str(best_alpha) + ' e número de neurônios: ' + str(best_number_centers)) plt.xlabel('Valor Esperado') plt.ylabel('Valor Encontrado') path = get_project_root() + '/run/TR-06/COLUNA_3C/results/' plt.savefig(path + "mat_confsuison_rbf.jpg") plt.show() print(pd.DataFrame(final_result)) del final_result['best_cf'] pd.DataFrame(final_result).to_csv(get_project_root() + '/run/TR-06/COLUNA_3C/results/' + 'result_rbf.csv')
# normalizar a base base[['x1', 'x2']] = normalization(base[['x1', 'x2']], type='min-max') x = array(base[['x1', 'x2']]) y = array(base[['y']]) classe0 = x[np.where(y == 0)[0]] classe1 = x[np.where(y == 1)[0]] classe2 = x[np.where(y == 2)[0]] plt.plot(classe0[:, 0], classe0[:, 1], 'b^') plt.plot(classe1[:, 0], classe1[:, 1], 'go') plt.plot(classe2[:, 0], classe2[:, 1], 'm*') plt.xlabel("X1") plt.ylabel("X2") plt.savefig(get_project_root() + '/run/TR-03/ARTIFICIAL/results/' + 'dataset_artificial.png') plt.show() y_out_of_c = pd.get_dummies(base['y']) base = base.drop(['y'], axis=1) base = concatenate([base[['x1', 'x2']], y_out_of_c], axis=1) for realization in range(20): train, test = split_random(base, train_percentage=.8) train, train_val = split_random(train, train_percentage=.8) x_train = train[:, :2] y_train = train[:, 2:]
base = pd.DataFrame(load_mock(type='LOGICAL_XOR'), columns=['x1', 'x2', 'y']) base[['x1', 'x2']] = normalization(base[['x1', 'x2']], type='min-max') x = array(base[['x1', 'x2']]) y = array(base[['y']]) classe0 = x[np.where(y == 0)[0]] classe1 = x[np.where(y == 1)[0]] plt.plot(classe0[:, 0], classe0[:, 1], 'b^') plt.plot(classe1[:, 0], classe1[:, 1], 'go') plt.xlabel("X1") plt.ylabel("X2") plt.savefig(get_project_root() + '/run/TR-05/XOR/results/' + 'dataset_xor_artificial.png') plt.show() # ----------------------- one - hot --------------------------------------------------- N, M = base.shape C = len(base['y'].unique()) y_out_of_c = pd.get_dummies(base['y']) base = concatenate([base[['x1', 'x2']], y_out_of_c], axis=1) # -------------------------------------------------------------------------------------- for realization in range(20): train, test = split_random(base, train_percentage=.8) train, train_val = split_random(train, train_percentage=.8)
train, test = split_random(base, train_percentage=.8) train = train.to_numpy() test = test.to_numpy() x_train = train[:, :len(features)] y_train = train[:, len(features):] x_test = test[:, :len(features)] y_test = test[:, len(features):] classifier_dmc = dmc(x_train, y_train) y_out_dmc = classifier_dmc.predict(x_test, [0, 1]) metrics_calculator = metric(list(y_test.reshape(y_test.shape[0])), y_out_dmc, types=['ACCURACY', 'AUC', 'precision', 'recall', 'f1_score']) metric_results = metrics_calculator.calculate(average='micro') results['realization'].append(realization) for type in ['ACCURACY', 'precision', 'recall', 'f1_score']: results[type].append(metric_results[type]) for type in ['ACCURACY', 'precision', 'recall', 'f1_score']: final_result[type].append(mean(results[type])) final_result['std ' + type].append(std(results[type])) print(DataFrame(final_result)) DataFrame(final_result).to_csv(get_project_root() + '/run/ML-00/COLUNA_2C/results/' + 'result_dmc.csv')
final_result['best_cf'].append(results['cf'][0][1]) final_result['ErrosxEpocohs'].append(results['erros'][0][1]) final_result['versus'].append(one_versus_others) final_result['alphas'].append(mean(results['alphas'])) for type in ['ACCURACY', 'AUC', 'precision', 'recall', 'f1_score']: final_result[type].append(mean(results[type])) final_result['std ' + type].append(std(results[type])) for i in range(len(final_result['best_cf'])): plt.figure(figsize=(10, 7)) df_cm = DataFrame(final_result['best_cf'][i], index=[i for i in "01"], columns=[i for i in "01"]) sn.heatmap(df_cm, annot=True) path = get_project_root() + '/run/TR-035/IRIS/results/' plt.savefig(path + 'mat_confsuison_hyper' + final_result['versus'][i] + ".jpg") plt.show() # for i in range(len(final_result['ErrosxEpocohs'])): # plt.plot(list(range(len(final_result['ErrosxEpocohs'][i]))), final_result['ErrosxEpocohs'][i], # label='Error') # plt.xlabel('Epocas') # plt.ylabel('Error') # plt.legend(loc='upper right') # # path = get_project_root() + '/run/TR-01/IRIS/results/' # plt.savefig(path + 'EpochsXError ' + final_result['versus'][i] + ".jpg") # plt.show() # #
results['error_per_epoch'].append( (simple_net.train_epochs_error, metric_results['RMSE'])) results['alphas'].append(simple_net.lr) results['realization'].append(realization) for type in ['MSE', 'RMSE', 'R2']: results[type].append(metric_results[type]) results['error_per_epoch'].sort(key=lambda x: x[1], reverse=False) final_result['best_error_per_epoch'] = results['error_per_epoch'][0][0] final_result['alphas'].append(mean(results['alphas'])) for type in ['MSE', 'RMSE', 'R2']: final_result[type].append(mean(results[type])) final_result['std ' + type].append(std(results[type])) # print(pd.DataFrame(final_result)) plt.plot(list(range(epochs)), final_result['best_error_per_epoch'], '*') plt.xlabel('epochs') plt.ylabel('RMSE') path = get_project_root() + '/run/TR-05/ABALONE/results/' plt.savefig(path + "error_epochs.jpg") plt.show() del final_result['best_error_per_epoch'] print(pd.DataFrame(final_result)) pd.DataFrame(final_result).to_csv(get_project_root() + '/run/TR-05/ABALONE/results/' + 'result_simple_net.csv') # ------------------------------------------------------------------------------------
simple_net = ExtremeLearningMachines(number_of_neurons=12, N_Classes=1, case='regression') simple_net.fit(x_train, y_train, x_train_val=x_train_val, y_train_val=y_train_val, hidden=number_of_neurons) y_out = simple_net.predict(x_test, bias=True) metrics_calculator = metric(y_test, y_out, types=['MSE', 'RMSE', 'R2']) metric_results = metrics_calculator.calculate() print(metric_results) results['realization'].append(realization) for type in ['MSE', 'RMSE', 'R2']: results[type].append(metric_results[type]) for type in ['MSE', 'RMSE', 'R2']: final_result[type].append(mean(results[type])) final_result['std ' + type].append(std(results[type])) print(pd.DataFrame(final_result)) pd.DataFrame(final_result).to_csv( get_project_root() + '/run/TR-06/ABALONE/results/' + 'result_elm.csv') # ------------------------------------------------------------------------------------
best_number_centers = results['cf'][0][3] final_result['alphas'].append(mean(results['alphas'])) for type in ['ACCURACY', 'precision', 'recall', 'f1_score']: final_result[type].append(mean(results[type])) final_result['std ' + type].append(std(results[type])) # ------------------------ PLOT ------------------------------------------------- for i in range(len(final_result['best_cf'])): plt.figure(figsize=(10, 7)) df_cm = DataFrame(final_result['best_cf'][i], index=[i for i in [0, 1, 2]], columns=[i for i in [0, 1, 2]]) sn.heatmap(df_cm, annot=True) plt.title( 'Matriz de connfusão dermatologia com raio de abertura: ' + str( best_alpha) + ' e numero de neurônios: ' + str(best_number_centers)) plt.xlabel('Valor Esperado') plt.ylabel('Valor Encontrado') path = get_project_root() + '/run/TR-06/IRIS/results/' plt.savefig(path + "mat_confsuison_iris_rbf.jpg") plt.show() print(pd.DataFrame(final_result)) del final_result['best_cf'] pd.DataFrame(final_result).to_csv(get_project_root() + '/run/TR-06/IRIS/results/' + 'result_rbf.csv')
} marker = { 0: '^', 1: 'o', } # O clasificador da vigesima realização plot_dict = { 'xx': xx, 'yy': yy, 'Z': classifier_perceptron.predict(space), 'classes': {} } # utilizando o x_test e o y_test da ultima realização for c in [0, 1]: plot_dict['classes'].update({ c: { 'X': x[where(y == c)[0]], 'point': point[c], 'marker': marker[c] } }) path = get_project_root() + '/run/TR-035/IRIS/results/' + 'color_map_' + str(combination) + str(one_versus_others) + '.jpg' coloring(plot_dict, ListedColormap(['#87CEFA', '#228B22']), xlabel=combination[0], ylabel=combination[1], title='mapa de cores com Rede Perceptron' + str(one_versus_others), xlim=[-0.1, 1.1], ylim=[-0.1, 1.1], path=path, save=True)
results['cf'].sort(key=lambda x: x[0], reverse=True) final_result['best_cf'].append(results['cf'][0][1]) best_number_neurons = results['cf'][0][2] for type in ['ACCURACY', 'precision', 'recall', 'f1_score']: final_result[type].append(mean(results[type])) final_result['std ' + type].append(std(results[type])) # ------------------------ PLOT ------------------------------------------------- # for i in range(len(final_result['best_cf'])): plt.figure(figsize=(10, 7)) df_cm = DataFrame(final_result['best_cf'][i], index=[i for i in range(C)], columns=[i for i in range(C)]) sn.heatmap(df_cm, annot=True) plt.title('Matriz de connfusão dermatologia com número de neurônios: ' + str(best_number_neurons)) plt.xlabel('Valor Esperado') plt.ylabel('Valor Encontrado') path = get_project_root() + '/run/TR-06/CANCER/results/' plt.savefig(path + "mat_confsuison_elm.jpg") plt.show() print(pd.DataFrame(final_result)) # del final_result['best_cf'] pd.DataFrame(final_result).to_csv(get_project_root() + '/run/TR-06/CANCER/results/' + 'result_elm.csv')
2: '*' } # O clasificador da vigesima realização plot_dict = { 'xx': xx, 'yy': yy, 'Z': out_of_c_to_label(simple_net.predict(space)), 'classes': {} } # utilizando o x_test e o y_test da ultima realização for c in [0, 1, 2]: plot_dict['classes'].update({ c: { 'X': x[where(y == c)[0]], 'point': point[c], 'marker': marker[c] } }) # #FFAAAA red # #AAAAFF blue coloring(plot_dict, ListedColormap(['#87CEFA', '#228B22', "#FF00FF"]), xlabel='SepalLengthCm', ylabel='SepalWidthCm', title='mapa de cores com Rede Perceptron', xlim=[-0.1, 1.1], ylim=[-0.1, 1.1], path=get_project_root() + '/run/TR-03/IRIS/results/' + 'color_map_sepal.jpg', save=True) # print('dataset shape %s' % Counter(base[:, 2]))
alphas=validation_alphas, hidden=hidden, validation=False) y_out = simple_net.predict(x_test, bias=True) metrics_calculator = metric(y_test, y_out, types=['MSE', 'RMSE']) metric_results = metrics_calculator.calculate() print(metric_results) results['alphas'].append(simple_net.lr) results['realization'].append(realization) for type in ['MSE', 'RMSE']: results[type].append(metric_results[type]) final_result['alphas'].append(mean(results['alphas'])) for type in ['MSE', 'RMSE']: final_result[type].append(mean(results[type])) final_result['std ' + type].append(std(results[type])) print(pd.DataFrame(final_result)) pd.DataFrame(final_result).to_csv( get_project_root() + '/run/TR-05/ELETRIC_MOTOR_TEMPERATURE/results/' + 'result_mlp_' + different_target + '.csv') # ------------------------------------------------------------------------------------
base = pd.DataFrame(load_mock(type='LOGICAL_XOR'), columns=['x1', 'x2', 'y']) base[['x1', 'x2']] = normalization(base[['x1', 'x2']], type='min-max') x = array(base[['x1', 'x2']]) y = array(base[['y']]) classe0 = x[np.where(y == 0)[0]] classe1 = x[np.where(y == 1)[0]] plt.plot(classe0[:, 0], classe0[:, 1], 'b^') plt.plot(classe1[:, 0], classe1[:, 1], 'go') plt.xlabel("X1") plt.ylabel("X2") plt.savefig(get_project_root() + '/run/TR-05/XOR/results/' + 'dataset_xor_artificial.png') plt.show() # ----------------------- one - hot --------------------------------------------------- N, M = base.shape C = len(base['y'].unique()) y_out_of_c = pd.get_dummies(base['y']) base = base.to_numpy() # -------------------------------------------------------------------------------------- for realization in range(1): train, test = split_random(base, train_percentage=.8)
best_number_centers = results['cf'][0][3] final_result['alphas'].append(mean(results['alphas'])) for type in ['ACCURACY', 'precision', 'recall', 'f1_score']: final_result[type].append(mean(results[type])) final_result['std ' + type].append(std(results[type])) # ------------------------ PLOT ------------------------------------------------- for i in range(len(final_result['best_cf'])): plt.figure(figsize=(10, 7)) df_cm = DataFrame(final_result['best_cf'][i], index=[i for i in range(C)], columns=[i for i in range(C)]) sn.heatmap(df_cm, annot=True) plt.title( 'Matriz de connfusão dermatologia com raio de abertura: ' + str(best_alpha) + ' e número de neurônios: ' + str( best_number_centers)) plt.xlabel('Valor Esperado') plt.ylabel('Valor Encontrado') path = get_project_root() + '/run/TR-06/COLUNA_2C/results/' plt.savefig(path + "mat_confsuison_rbf.jpg") plt.show() print(pd.DataFrame(final_result)) del final_result['best_cf'] pd.DataFrame(final_result).to_csv(get_project_root() + '/run/TR-06/COLUNA_2C/results/' + 'result_rbf_net.csv')
best_number_neurons = results['cf'][0][2] for type in ['ACCURACY', 'precision', 'recall', 'f1_score']: final_result[type].append(mean(results[type])) final_result['std ' + type].append(std(results[type])) # ------------------------ PLOT ------------------------------------------------- for i in range(len(final_result['best_cf'])): plt.figure(figsize=(10, 7)) df_cm = DataFrame(final_result['best_cf'][i], index=[i for i in range(C)], columns=[i for i in range(C)]) sn.heatmap(df_cm, annot=True) plt.title( 'Matriz de connfusão dermatologia com número de neurônios: ' + str(best_number_neurons)) plt.xlabel('Valor Esperado') plt.ylabel('Valor Encontrado') path = get_project_root() + '/run/TR-06/DERMATOLOGIA/results/' plt.savefig(path + "mat_confsuison_elm.jpg") plt.show() print(pd.DataFrame(final_result)) del final_result['best_cf'] pd.DataFrame(final_result).to_csv(get_project_root() + '/run/TR-06/DERMATOLOGIA/results/' + 'result_elm.csv')
results[type].append(metric_results[type]) final_result['versus'].append(one_versus_others) final_result['K'].append(3) for type in ['ACCURACY', 'precision', 'recall', 'f1_score']: final_result[type].append(mean(results[type])) final_result['std ' + type].append(std(results[type])) results['cf'].sort(key=lambda x: x[0], reverse=True) final_result['best_cf'].append(results['cf'][0][1]) best_acc_clf = results['cf'][0][2] best_acc = results['cf'][0][0] df_cm = DataFrame(results['cf'][0][1], index=[i for i in range(C)], columns=[i for i in range(C)]) sns.heatmap(df_cm, annot=True) plt.title('Matriz de connfusão do KNN com acurácia de ' + str(round(best_acc, 2) * 100) + "%") plt.xlabel('Valor Esperado') plt.ylabel('Valor Encontrado') path = get_project_root() + '/run/TR-00/IRIS/results/' plt.savefig(path + one_versus_others + "-conf_result_knn.jpg") plt.show() DataFrame(final_result).to_csv(get_project_root() + '/run/TR-00/IRIS/results/' + 'result_knn.csv') print(DataFrame(final_result))
y_train, x_train_val=x_train_val, y_train_val=y_train_val, alphas=validation_alphas, hidden=hidden, validation=False) y_out = simple_net.predict(x_test, bias=True) metrics_calculator = metric(y_test, y_out, types=['MSE', 'RMSE', 'R2']) metric_results = metrics_calculator.calculate() print(metric_results) results['realization'].append(realization) for type in ['MSE', 'RMSE', 'R2']: results[type].append(metric_results[type]) for type in ['MSE', 'RMSE', 'R2']: final_result[type].append(mean(results[type])) final_result['std ' + type].append(std(results[type])) print(pd.DataFrame(final_result)) pd.DataFrame(final_result).to_csv( get_project_root() + '/run/TR-06/ELETRIC_MOTOR_TEMPERATURE/results/' + 'result_rbf_' + different_target + '.csv') # ------------------------------------------------------------------------------------
print(metric_results) results['alphas'].append(simple_net.learning_rate) results['realization'].append(realization) for type in ['ACCURACY', 'precision', 'recall', 'f1_score']: results[type].append(metric_results[type]) final_result['alphas'].append(mean(results['alphas'])) for type in ['ACCURACY', 'precision', 'recall', 'f1_score']: final_result[type].append(mean(results[type])) final_result['std ' + type].append(std(results[type])) # ------------------------ PLOT ------------------------------------------------- # for i in range(len(final_result['best_cf'])): # plt.figure(figsize=(10, 7)) # # df_cm = DataFrame(final_result['best_cf'][i], index=[i for i in "012"], # columns=[i for i in "012"]) # sn.heatmap(df_cm, annot=True) # # path = get_project_root() + '/run/TR-03/ARTIFICIAL/results/' # plt.savefig(path + "mat_confsuison_triangle.jpg") # plt.show() print(pd.DataFrame(final_result)) # del final_result['best_cf'] pd.DataFrame(final_result).to_csv(get_project_root() + '/run/TR-03/COLUNA_3C/results/' + 'result_simple_net.csv')
'MSE': [], 'RMSE': [], 'erros': [], 'alphas': [] } F, x1, x2 = load_mock(type='2D_REGRESSOR') fig = plt.figure() ax = fig.add_subplot(111, projection='3d') ax.scatter(x1, x2, F, color='k') ax.set_title("y = ax1 + bx2 + c") ax.set_xlabel("x1") ax.set_ylabel("x2") ax.set_zlabel("y") plt.savefig(get_project_root() + '/run/TR-02/ARTIFICIAL/results/' + 'adaline_fig_2.jpg') plt.show() base = concatenate( [array(x1, ndmin=2), array(x2, ndmin=2), array(F, ndmin=2)], axis=0).T validation_alphas = linspace(0.015, 0.1, 20) for realization in range(20): print("Realization" + str(realization)) train, test = split_random(base, train_percentage=.8) train, train_val = split_random(train, train_percentage=0.7) x_train = train[:, :2]
print(metric_results) results['alphas'].append(simple_net.learning_rate) results['realization'].append(realization) for type in ['ACCURACY', 'precision', 'recall', 'f1_score']: results[type].append(metric_results[type]) final_result['alphas'].append(mean(results['alphas'])) for type in ['ACCURACY', 'precision', 'recall', 'f1_score']: final_result[type].append(mean(results[type])) final_result['std ' + type].append(std(results[type])) # ------------------------ PLOT ------------------------------------------------- # for i in range(len(final_result['best_cf'])): # plt.figure(figsize=(10, 7)) # # df_cm = DataFrame(final_result['best_cf'][i], index=[i for i in ['']], # columns=[i for i in ['']]) # sn.heatmap(df_cm, annot=True) # # path = get_project_root() + '/run/TR-03/IRIS/results/' # plt.savefig(path + "mat_confsuison_iris.jpg") # plt.show() print(pd.DataFrame(final_result)) # del final_result['best_cf'] pd.DataFrame(final_result).to_csv(get_project_root() + '/run/TR-03/IRIS/results/' + 'result_simple_net.csv')
results['cf'].sort(key=lambda x: x[0], reverse=True) final_result['best_cf'].append(results['cf'][0][1]) best_number_neurons = results['cf'][0][2] for type in ['ACCURACY', 'precision', 'recall', 'f1_score']: final_result[type].append(mean(results[type])) final_result['std ' + type].append(std(results[type])) # ------------------------ PLOT ------------------------------------------------- for i in range(len(final_result['best_cf'])): plt.figure(figsize=(10, 7)) df_cm = DataFrame(final_result['best_cf'][i], index=[i for i in range(C)], columns=[i for i in range(C)]) sn.heatmap(df_cm, annot=True) plt.title( 'Matriz de connfusão dermatologia com número de neurônios: ' + str(best_number_neurons)) plt.xlabel('Valor Esperado') plt.ylabel('Valor Encontrado') path = get_project_root() + '/run/TR-06/COLUNA_3C/results/' plt.savefig(path + "mat_confsuison_elm.jpg") plt.show() # del final_result['best_cf'] # print(pd.DataFrame(final_result)) # pd.DataFrame(final_result).to_csv(get_project_root() + '/run/TR-06/COLUNA_3C/results/' + 'result_elm.csv')
final_result[type].append(mean(results[type])) final_result['std ' + type].append(std(results[type])) results['cf'].sort(key=lambda x: x[0], reverse=True) final_result['best_cf'].append(results['cf'][0][1]) best_acc_clf = results['cf'][0][2] best_acc = results['cf'][0][0] df_cm = DataFrame(results['cf'][0][1], index=[i for i in range(C)], columns=[i for i in range(C)]) sn.heatmap(df_cm, annot=True) plt.title('Matriz de connfusão do KNN com acurácia de ' + str(round(best_acc, 2) * 100) + "%") plt.xlabel('Valor Esperado') plt.ylabel('Valor Encontrado') path = get_project_root() + '/run/ML-00/COLUNA_3C/results/' plt.savefig(path + one_versus_others + "-conf_result_knn.jpg") plt.show() print(DataFrame(final_result)) DataFrame(final_result).to_csv(get_project_root() + '/run/ML-00/COLUNA_3C/results/' + one_versus_others + '_result_knn.csv') xx, yy = generate_space(x_test) space = c_[xx.ravel(), yy.ravel()] point = { 0: 'ro', 1: 'bo' } marker = { 0: 's',
results['alphas'].append(simple_net.learning_rate) results['realization'].append(realization) for type in ['ACCURACY', 'precision', 'recall', 'f1_score']: results[type].append(metric_results[type]) results['cf'].sort(key=lambda x: x[0], reverse=True) final_result['best_cf'].append(results['cf'][0][1]) final_result['alphas'].append(mean(results['alphas'])) for type in ['ACCURACY', 'precision', 'recall', 'f1_score']: final_result[type].append(mean(results[type])) final_result['std ' + type].append(std(results[type])) # ------------------------ PLOT ------------------------------------------------- for i in range(len(final_result['best_cf'])): plt.figure(figsize=(10, 7)) df_cm = DataFrame(final_result['best_cf'][i], index=[i for i in range(C)], columns=[i for i in range(C)]) sn.heatmap(df_cm, annot=True) path = get_project_root() + '/run/TR-05/DERMATOLOGIA/results/' plt.savefig(path + "mat_confsuison.jpg") plt.show() # print(pd.DataFrame(final_result)) # del final_result['best_cf'] # pd.DataFrame(final_result).to_csv(get_project_root() + '/run/TR-05/DERMATOLOGIA/results/' + 'result_mlp.csv')
results['realization'].append(realization) for type in ['ACCURACY', 'precision', 'recall', 'f1_score']: results[type].append(metric_results[type]) results['cf'].sort(key=lambda x: x[0], reverse=True) final_result['best_cf'].append(results['cf'][0][1]) final_result['alphas'].append(mean(results['alphas'])) for type in ['ACCURACY', 'precision', 'recall', 'f1_score']: final_result[type].append(mean(results[type])) final_result['std ' + type].append(std(results[type])) # ------------------------ PLOT ------------------------------------------------- for i in range(len(final_result['best_cf'])): plt.figure(figsize=(10, 7)) df_cm = DataFrame(final_result['best_cf'][i], index=[i for i in [0,1,2]], columns=[i for i in [0,1,2]]) sn.heatmap(df_cm, annot=True) path = get_project_root() + '/run/TR-05/IRIS/results/' plt.savefig(path + "mat_confsuison_iris.jpg") plt.show() print(pd.DataFrame(final_result)) # del final_result['best_cf'] # pd.DataFrame(final_result).to_csv(get_project_root() + '/run/TR-05/IRIS/results/' + 'result_mlp.csv')
'std RMSE': [], 'R2': [], 'std R2': [] } results = {'realization': [], 'MSE': [], 'RMSE': [], 'R2': []} base = load_mock(type='MOCK_SENO') # normalizar a base base[['x1']] = normalization(base[['x1']], type='min-max') sn.set_style('whitegrid') sn.scatterplot(data=base, x="x1", y="y", color='c') plt.xlabel("X1") plt.ylabel("Y") plt.savefig(get_project_root() + '/run/TR-05/ARTIFICIAL_REGRESSAO/results/' + 'dataset_seno_artificial.png') plt.show() x = array(base[['x1']]) y = array(base[['y']]) N, M = base.shape C = 1 # Problema de regressão epochs = 2000 for realization in range(20): train, test = split_random(base, train_percentage=.8) train, train_val = split_random(train, train_percentage=.8)