#testing on the training data through k-cross-fold validation tester.startClassificationTest() # starting on the evaluation set X_train = X_train_origin[features] X_test = X_test_origin[features] i = 0 result = Results() #getting test results for each classifier while (i < len(clfs)): clfs[i].fit(X_train, Y_train) preds = clfs[i].predict(X_test) result.accuracy = metrics.accuracy_score(Y_test, preds) result.precision = metrics.precision_score(Y_test, preds) result.recall = metrics.recall_score(Y_test, preds) result.k_cohen = metrics.cohen_kappa_score(Y_test, preds) result.f1_measure = metrics.f1_score(Y_test, preds) result.log_loss = metrics.log_loss(Y_test, clfs[i].predict_proba(X_test)) #write results into file printResults(result, clfNames[i], len(features)) i += 1 featureSize -= 5 #plotting test and train results dirPath = "Classification/Test/" plotter = Plotter(clfNames, dirPath) metricNames = ["Accuracy", "Precision", "Recall", "K_cohen", "F1_measure", "Log-loss"]
#testing on the training data through k-cross-fold validation tester.startClassificationTest() # starting on the evaluation set X_train = X_train_origin[features] X_test = X_test_origin[features] i = 0 result = Results() #getting test results for each classifier while (i < len(clfs)): clfs[i].fit(X_train, Y_train) preds = clfs[i].predict(X_test) result.accuracy = metrics.accuracy_score(Y_test, preds) result.precision = metrics.precision_score(Y_test, preds) result.recall = metrics.recall_score(Y_test, preds) result.k_cohen = metrics.cohen_kappa_score(Y_test, preds) result.f1_measure = metrics.f1_score(Y_test, preds) result.log_loss = metrics.log_loss(Y_test, clfs[i].predict_proba(X_test)) #write results into file printResults(result, clfNames[i], len(features)) i += 1 featureSize -= 5 #plotting test and train results dirPath = "Classification/Test/" plotter = Plotter(clfNames, dirPath)