def run_argument_sets(text_clf, argument_sets):
scores = []
for s in argument_sets:
(X_train, X_test, y_train, y_test) = s
print("---Fitting model---")
text_clf.fit(X_train, y_train)
joblib.dump(text_clf, 'svm_trained.joblib')
print("SVM with SGD")
documents_predicted = []
documents_target = []
all_predicted_lines = []
all_target_lines = []
for doc in doc_test:
predicted_lines = text_clf.predict(doc.data)
all_predicted_lines += list(predicted_lines)
all_target_lines += list(doc.targets)
predicted_doc = utils.classify_doc(predicted_lines)
documents_predicted.append(predicted_doc)
documents_target.append(doc.category)
scores += [text_clf.score(X_test, y_test)]
print("Line by Line ")
print("Confusion Matrix: \n{}".format(
confusion_matrix(all_target_lines, all_predicted_lines)))
accuracy = fbeta_score(all_target_lines,
all_predicted_lines,
average=None,
beta=2)
print("Accuracy: {}".format(accuracy))
print("Scores: ", scores)
print("Scores:", np.mean(scores))
shuffle=True))])
print("Training Model")
text_clf.fit(X_train, y_train)
print("SGD")
joblib.dump(text_clf, 'svm_trained.joblib')
documents_predicted = []
documents_target = []
all_predicted_lines = []
all_target_lines = []
for doc in doc_test:
predicted_lines = text_clf.predict(doc.data)
all_predicted_lines += list(predicted_lines)
all_target_lines += list(doc.targets)
predicted_doc = utils.classify_doc(predicted_lines)
documents_predicted.append(predicted_doc)
documents_target.append(doc.category)
print("Line by Line ")
print("Confusion Matrix: \n{}".format(
confusion_matrix(all_target_lines, all_predicted_lines)))
accuracy = fbeta_score(all_target_lines,
all_predicted_lines,
average=None,
beta=2)
print("Accuracy: {}".format(accuracy))
doc_accuracy = fbeta_score(documents_target,
documents_predicted,