def main():
"""
Main function.
:return:
"""
X, Y = getFeatureLabelData(train_url_cancer, 1)
#If we want to test scaled data uncomment X_values and replace x.values with x_values
#X_values = minMaxScailing(X)
X_sampled, Y_sampled = underSampling(X, Y)
x_train, x_test, y_train, y_test = trainTestSplit(X_sampled, Y_sampled)
#x_train, x_test, y_train, y_test = trainTestSplit(X.values, Y)
#if we want to normalize data remove next comment
x_train = minMaxScailing(x_train)
x_test = minMaxScailing(x_test)
print(
"-----------------------Feature Selection-----------------------------------------"
)
print(x_test)
x_train, x_test = selectRandomForests(x_train, y_train, x_test, y_test)
print(
"-----------------------After Selection-----------------------------------------"
)
print(x_test)
print(
"-----------------------NaiveBayes-----------------------------------------"
)
nb_label_prediction, proba = getPredictionData("NaiveBayes", x_train,
x_test, y_train, y_test)
print(confusion_matrix_results(y_test, nb_label_prediction))
print(classification_report_results(y_test, nb_label_prediction))
print(classification_accuracy_score(y_test, nb_label_prediction))
print(classification_roc_auc_score(y_test, nb_label_prediction, proba, 0))
print(
"----------------------------kNeighbours------------------------------------------"
)
kn_label_prediction, proba = getPredictionData("kNeighbours", x_train,
x_test, y_train, y_test, 5)
print(confusion_matrix_results(y_test, kn_label_prediction))
print(classification_report_results(y_test, kn_label_prediction))
print(classification_accuracy_score(y_test, kn_label_prediction))
print(classification_roc_auc_score(y_test, kn_label_prediction, proba, 0))
print(
"----------------------------decisionTree------------------------------------------"
)
dt_label_prediction, proba = getPredictionData("decisionTree", x_train,
x_test, y_train, y_test)
print(confusion_matrix_results(y_test, dt_label_prediction))
print(classification_report_results(y_test, dt_label_prediction))
print(classification_accuracy_score(y_test, dt_label_prediction))
print(classification_roc_auc_score(y_test, dt_label_prediction, proba, 0))
def bank_test(samplingType,
oneHotEncoding,
normalizeType,
selectBest,
kBest=10):
print(
"----------------------------Preprocessing------------------------------------------"
)
start_preprocessing_time = time.time()
dataBank = pandas.read_csv('../data/Bank_Marketing_Data/bank-full.csv',
sep=";")
X = dataBank.drop(dataBank.columns[[10, -1]], axis=1)
Y = dataBank.iloc[:, -1].values
X = featureEncoding(X.values, [1, 2, 3, 4, 6, 7, 8, 9, 10, 14])
if (samplingType == 'over'):
X_sampled, Y_sampled = overSampling(X, Y)
x_train, x_test, y_train, y_test = trainTestSplit(X_sampled, Y_sampled)
elif (samplingType == 'under'):
X_sampled, Y_sampled = underSampling(X, Y)
x_train, x_test, y_train, y_test = trainTestSplit(X_sampled, Y_sampled)
elif (samplingType == 'combined'):
X_sampled, Y_sampled = combinedSampling(X, Y)
x_train, x_test, y_train, y_test = trainTestSplit(X_sampled, Y_sampled)
else:
x_train, x_test, y_train, y_test = trainTestSplit(X, Y)
if oneHotEncoding:
x_train, x_test = featureOneHotEncoding(x_train, x_test)
#Pre processing to be added here, not after the preprocessing_time
if (normalizeType == 'minmax'):
x_train = minMaxScailing(x_train)
x_test = minMaxScailing(x_test)
elif (normalizeType == 'normal'):
x_train, x_test = normalizeData(x_train, x_test)
if selectBest:
x_train, x_test = selectKBest(x_train, y_train, x_test, y_test, kBest)
preprocessing_time = time.time()
print("The processing time is: ",
preprocessing_time - start_preprocessing_time)
print(
"----------------------------end Preprocessing------------------------------------------"
)
print(
"----------------------------decisionTree------------------------------------------"
)
start_prediction_dt_time = time.time()
dt_label_prediction, proba = getPredictionData("decisionTree", x_train,
x_test, y_train, y_test)
prediction_dt_time = time.time()
print(confusion_matrix_results(y_test, dt_label_prediction))
print(classification_report_results(y_test, dt_label_prediction))
print("The accuracy is: ",
classification_accuracy_score(y_test, dt_label_prediction))
print('The auc_roc score is: ',
classification_roc_auc_score(y_test, dt_label_prediction, proba, 0))
print("The prediction time for Decision tree is:",
prediction_dt_time - start_prediction_dt_time)
print(
"----------------------------end decisionTree------------------------------------------"
)
print(
"-----------------------NaiveBayes-----------------------------------------"
)
start_prediction_bayes_time = time.time()
nb_label_prediction, proba = getPredictionData("NaiveBayes", x_train,
x_test, y_train, y_test)
prediction_bayes_time = time.time()
print(confusion_matrix_results(y_test, nb_label_prediction))
print(classification_report_results(y_test, nb_label_prediction))
print("The accuracy is: ",
classification_accuracy_score(y_test, nb_label_prediction))
print('The auc_roc score is: ',
classification_roc_auc_score(y_test, nb_label_prediction, proba, 0))
print("The prediction time for Naive Bayes is:",
prediction_bayes_time - start_prediction_bayes_time)
print(
"-----------------------End NaiveBayes-----------------------------------------"
)
print(
"----------------------------kNeighbours------------------------------------------"
)
start_prediction_knn_time = time.time()
kn_label_prediction, proba = getPredictionData("kNeighbours", x_train,
x_test, y_train, y_test, 5)
prediction_knn_time = time.time()
print(confusion_matrix_results(y_test, kn_label_prediction))
print(classification_report_results(y_test, kn_label_prediction))
print('The accuracy is: ',
classification_accuracy_score(y_test, kn_label_prediction))
print('The auc_roc score is: ',
classification_roc_auc_score(y_test, kn_label_prediction, proba, 0))
print("The prediction time for kNN is:",
prediction_knn_time - start_prediction_knn_time)
print(
"----------------------------End kNeighbours------------------------------------------"
)
def amazon_measure(scaling=False, sampling="None", featureSelection=False):
"""
Main function.
:return:
"""
print(
"----------------------------Preprocessing------------------------------------------"
)
start_preprocessing_time = time.time()
train_url = '../data/Amazon_Review_Data/amazon_review_ID.shuf.lrn.csv'
X, Y = getFeatureLabelData(train_url, -1)
if sampling == "Under":
X_sampled, Y_sampled = underSampling(X, Y)
elif sampling == "Over":
X_sampled, Y_sampled = overSampling(X, Y)
elif sampling == "Combined":
X_sampled, Y_sampled = combinedSampling(X, Y)
else:
X_sampled, Y_sampled = X.values, Y
if scaling:
X_values = minMaxScailing(X_sampled)
else:
X_values = X_sampled
x_train, x_test, y_train, y_test = trainTestSplit(X_sampled, Y_sampled)
if featureSelection:
x_train, x_test = selectKBest(x_train, y_train, x_test, y_test)
# Pre processing to be added here, not after the preprocessing_time
preprocessing_time = time.time()
print("The processing time is: ",
preprocessing_time - start_preprocessing_time)
print(
"----------------------------end Preprocessing------------------------------------------"
)
print(
"----------------------------decisionTree------------------------------------------"
)
start_prediction_dt_time = time.time()
dt_label_prediction, proba = getPredictionData("decisionTree", x_train,
x_test, y_train, y_test)
prediction_dt_time = time.time()
print(confusion_matrix_results(y_test, dt_label_prediction))
print(classification_report_results(y_test, dt_label_prediction))
print("The accuracy is: ",
classification_accuracy_score(y_test, dt_label_prediction))
#print('The auc_roc score is: ', classification_roc_auc_score(y_test, dt_label_prediction, proba, 0))
print("The prediction time for Decision tree is:",
prediction_dt_time - start_prediction_dt_time)
print(
"----------------------------end decisionTree------------------------------------------"
)
print(
"-----------------------NaiveBayes-----------------------------------------"
)
start_prediction_bayes_time = time.time()
nb_label_prediction, proba = getPredictionData("NaiveBayes", x_train,
x_test, y_train, y_test)
prediction_bayes_time = time.time()
print(confusion_matrix_results(y_test, nb_label_prediction))
print(classification_report_results(y_test, nb_label_prediction))
print("The accuracy is: ",
classification_accuracy_score(y_test, nb_label_prediction))
#print('The auc_roc score is: ', classification_roc_auc_score(y_test, nb_label_prediction, proba, 0))
print("The prediction time for Naive Bayes is:",
prediction_bayes_time - start_prediction_bayes_time)
print(
"-----------------------End NaiveBayes-----------------------------------------"
)
print(
"----------------------------kNeighbours------------------------------------------"
)
start_prediction_knn_time = time.time()
kn_label_prediction, proba = getPredictionData("kNeighbours", x_train,
x_test, y_train, y_test, 5)
prediction_knn_time = time.time()
print(confusion_matrix_results(y_test, kn_label_prediction))
print(classification_report_results(y_test, kn_label_prediction))
print('The accuracy is: ',
classification_accuracy_score(y_test, kn_label_prediction))
#print('The auc_roc score is: ', classification_roc_auc_score(y_test, kn_label_prediction, proba, 0))
print("The prediction time for kNN is:",
prediction_knn_time - start_prediction_knn_time)
print(
"----------------------------End kNeighbours------------------------------------------"
)
def measure_data(url,
classColumn,
noID,
samplingType,
normalizeType,
selectBest,
kBest=10):
print(
"----------------------------Preprocessing------------------------------------------"
)
start_preprocessing_time = time.time()
X, Y = getFeatureLabelData(url, classColumn, noID)
if (samplingType == 'over'):
X_sampled, Y_sampled = overSampling(X, Y)
x_train, x_test, y_train, y_test = trainTestSplit(X_sampled, Y_sampled)
elif (samplingType == 'under'):
X_sampled, Y_sampled = underSampling(X, Y)
x_train, x_test, y_train, y_test = trainTestSplit(X_sampled, Y_sampled)
elif (samplingType == 'combined'):
X_sampled, Y_sampled = combinedSampling(X, Y)
x_train, x_test, y_train, y_test = trainTestSplit(X_sampled, Y_sampled)
else:
x_train, x_test, y_train, y_test = trainTestSplit(X.values, Y)
#Pre processing to be added here, not after the preprocessing_time
if (normalizeType == 'minmax'):
x_train = minMaxScailing(x_train)
x_test = minMaxScailing(x_test)
elif (normalizeType == 'normal'):
x_train, x_test = normalizeData(x_train, x_test)
if selectBest:
x_train, x_test = selectKBest(x_train, y_train, x_test, y_test, kBest)
preprocessing_time = time.time()
print("The processing time is: ",
preprocessing_time - start_preprocessing_time)
print(
"----------------------------end Preprocessing------------------------------------------"
)
print(
"----------------------------decisionTree------------------------------------------"
)
start_prediction_dt_time = time.time()
dt_label_prediction, proba = getPredictionData("decisionTree", x_train,
x_test, y_train, y_test)
prediction_dt_time = time.time()
print(confusion_matrix_results(y_test, dt_label_prediction))
print(classification_report_results(y_test, dt_label_prediction))
print("The accuracy is: ",
classification_accuracy_score(y_test, dt_label_prediction))
#print('The auc_roc score is: ', classification_roc_auc_score(y_test, dt_label_prediction, proba, 0))
print("The prediction time for Decision tree is:",
prediction_dt_time - start_prediction_dt_time)
print(
"----------------------------end decisionTree------------------------------------------"
)
print(
"-----------------------NaiveBayes-----------------------------------------"
)
start_prediction_bayes_time = time.time()
nb_label_prediction, proba = getPredictionData("NaiveBayes", x_train,
x_test, y_train, y_test)
prediction_bayes_time = time.time()
print(confusion_matrix_results(y_test, nb_label_prediction))
print(classification_report_results(y_test, nb_label_prediction))
print("The accuracy is: ",
classification_accuracy_score(y_test, nb_label_prediction))
#print('The auc_roc score is: ', classification_roc_auc_score(y_test, nb_label_prediction, proba, 0))
print("The prediction time for Naive Bayes is:",
prediction_bayes_time - start_prediction_bayes_time)
print(
"-----------------------End NaiveBayes-----------------------------------------"
)
print(
"----------------------------kNeighbours------------------------------------------"
)
start_prediction_knn_time = time.time()
kn_label_prediction, proba = getPredictionData("kNeighbours", x_train,
x_test, y_train, y_test, 5)
prediction_knn_time = time.time()
print(confusion_matrix_results(y_test, kn_label_prediction))
print(classification_report_results(y_test, kn_label_prediction))
print('The accuracy is: ',
classification_accuracy_score(y_test, kn_label_prediction))
#print('The auc_roc score is: ', classification_roc_auc_score(y_test, kn_label_prediction, proba, 0))
print("The prediction time for kNN is:",
prediction_knn_time - start_prediction_knn_time)
print(
"----------------------------End kNeighbours------------------------------------------"
)