testDataFiles = [ 'ada_regress_test.csv', # 'knn_regress_test.csv', 'log_regress_test.csv', 'NN_test.csv', 'rf_regress_test.csv', # 'svm_regress_test.csv' ] model_size = 0.7 y_blend_train = data.loadTrainingBlend()['ylabels'] x_blend_train = np.zeros((len(trainingDataFiles), len(y_blend_train))) x_blend_train = np.transpose([loadModelOut(model, model_size) for model in trainingDataFiles]) x_all = len(data.allTest()['xlabels']) x_test = np.zeros((len(testDataFiles), x_all)) x_test = np.transpose([loadModelOut(model, 0) for model in testDataFiles]) # parameters = {'C': np.logspace(-4.0, 4.0, 20), # # 'kernel': ['rbf'], # 'kernel': ['linear', 'poly', 'rbf'], # # 'degree': [2] # 'degree': np.arange(0.0, 4.0, 1), # } parameters = {'C': np.logspace(-2, 2, 10), # 'solver' : ['newton-cg', 'lbfgs', 'liblinear'] }
print "Training score: " print nn_class.score(x1, y1) print "Test score : " print nn_class.score(x_23, y_23) # cross_val_scores = cross_validation.cross_val_score(estimator=nn_class,\ # X=x_train, y=y_train, cv=kf_total, n_jobs=1) # print "cross val scores: " # print cross_val_scores x_all_test = data.allTest()['xlabels'] f = open('nn_regress_test.csv', 'w+') f.write('Id,Prediction\n') y_test = nn_class.predict(x_all_test) for i in range(len(y_test)): f.write(str(i+1) + ',' + str(y_test[i]) + '\n') f.close() g = open('nn_regress_params.txt', 'w+') g.write(str(nn_class.best_estimator_.get_params())) g.close() x_all_train = data.allTrain()['xlabels'] h = open('nn_regress_train.csv', 'w+')