################ Cross Validation Hyperparametre Tuning ###############################
print("\nHYPERPARAMETRE TUNING")
hyperparams = {'C': [0.1, 1, 100, 1000], 'penalty': ['l2'], 'loss': ['hinge', 'squared_hinge'], 'max_iter': [1000, 2000]}
optimized_model = GridSearchCV(estimator=support_vector_classifier, param_grid=hyperparams,
n_jobs=1, cv=3, verbose=1, error_score=1)
optimized_model.fit(X_train, y_train)
print(">>>>> Optimized params")
print(optimized_model.best_params_)
cv_results = optimized_model.cv_results_
print(">>>>>> Display the top results of grid search cv")
results_dataframe = pd.DataFrame(cv_results).sort_values(by='rank_test_score')
display( results_dataframe.head() )
saveDataframe(results_dataframe,"IMDB Support Vector Classifier")
prediction = optimized_model.predict(X_test)
score = np.mean(prediction == y_test)
print("Using our training-dataset optimized Support Vector Machine model on the testing dataset for evaluating")
print("score = %f" % score)
################ Graph Reporting ###############################
# plot the confusion matrix
skplt.metrics.plot_confusion_matrix(y_test, predictions, normalize=False, figsize=(12, 8))
plt.show()
param_grid=hyperparams,
n_jobs=1,
cv=3,
verbose=1,
error_score=1)
optimized_model.fit(X_train, y_train)
print(">>>>> Optimized params")
print(optimized_model.best_params_)
cv_results = optimized_model.cv_results_
print(">>>>>> Display the top results of grid search cv")
results_dataframe = pd.DataFrame(cv_results).sort_values(by='rank_test_score')
display(results_dataframe.head())
saveDataframe(results_dataframe, "IMDB AdaBoost")
prediction = optimized_model.predict(X_test)
score = np.mean(prediction == y_test)
print(
"Using our training-dataset optimized adaboost model on the testing dataset for evaluating"
)
print("score = %f" % score)
################ Graph Reporting ###############################
# plot the confusion matrix
skplt.metrics.plot_confusion_matrix(y_test,
predictions,
normalize=False,
figsize=(12, 8))
param_grid=hyperparams,
n_jobs=1,
cv=3,
verbose=1,
error_score=1)
optimized_model.fit(X_train, y_train)
print(">>>>> Optimized params")
print(optimized_model.best_params_)
cv_results = optimized_model.cv_results_
print(">>>>>> Display the top results of grid search cv")
results_dataframe = pd.DataFrame(cv_results).sort_values(by='rank_test_score')
display(results_dataframe.head())
saveDataframe(results_dataframe, "IMDB Logistic Regression")
prediction = optimized_model.predict(X_test)
score = np.mean(prediction == y_test)
print(
"Using our training-dataset optimized logistic regression model on the testing dataset for evaluating"
)
print("score = %f" % score)
################ Graph Reporting ###############################
# plot the confusion matrix
skplt.metrics.plot_confusion_matrix(y_test,
predictions,
normalize=False,
figsize=(12, 8))
param_grid=hyperparams,
n_jobs=1,
cv=3,
verbose=1,
error_score=1)
optimized_model.fit(X_train, y_train)
print(">>>>> Optimized params")
print(optimized_model.best_params_)
cv_results = optimized_model.cv_results_
print(">>>>>> Display the top results of grid search cv")
results_dataframe = pd.DataFrame(cv_results).sort_values(by='rank_test_score')
display(results_dataframe.head())
saveDataframe(results_dataframe, "IMDB Multinomial Naive Bayes")
prediction = optimized_model.predict(X_test)
score = np.mean(prediction == y_test)
print(
"Using our training-dataset optimized multinomial naive bayes model on the testing dataset for evaluating"
)
print("score = %f" % score)
################ Graph Reporting ###############################
# plot the confusion matrix
skplt.metrics.plot_confusion_matrix(y_test,
predictions,
normalize=False,
figsize=(12, 8))
param_grid=hyperparams,
n_jobs=1,
cv=3,
verbose=1,
error_score=1)
optimized_model.fit(X_train, y_train)
print(">>>>> Optimized params")
print(optimized_model.best_params_)
cv_results = optimized_model.cv_results_
print(">>>>>> Display the top results of grid search cv")
results_dataframe = pd.DataFrame(cv_results).sort_values(by='rank_test_score')
display(results_dataframe.head())
saveDataframe(results_dataframe, "IMDB Decision Tree")
prediction = optimized_model.predict(X_test)
score = np.mean(prediction == y_test)
print(
"Using our training-dataset optimized decision tree model on the testing dataset for evaluating"
)
print("score = %f" % score)
################ Graph Reporting ###############################
# plot the confusion matrix
skplt.metrics.plot_confusion_matrix(y_test,
predictions,
normalize=False,
figsize=(12, 8))
param_grid=hyperparams,
n_jobs=1,
cv=3,
verbose=1,
error_score=1)
optimized_model.fit(X_train, y_train)
print(">>>>> Optimized params")
print(optimized_model.best_params_)
cv_results = optimized_model.cv_results_
print(">>>>>> Display the top results of grid search cv")
results_dataframe = pd.DataFrame(cv_results).sort_values(by='rank_test_score')
display(results_dataframe.head())
saveDataframe(results_dataframe, "IMDB Random Forest")
prediction = optimized_model.predict(X_test)
score = np.mean(prediction == y_test)
print(
"Using our training-dataset optimized Random Forest model on the testing dataset for evaluating"
)
print("score = %f" % score)
################ Graph Reporting ###############################
# plot the confusion matrix
skplt.metrics.plot_confusion_matrix(y_test,
predictions,
normalize=False,
figsize=(12, 8))