def main():
data = DataHelper.Data()
x_train, y_train, x_test, y_test, _ = data.loadData("hw2_data_1.txt", 2, 70)
epoch = 3
weights = np.ones(len(x_train)) / len(x_train)
adaboost = Adaboost(weights, epoch, DecisionRule())
adaboost.train(x_train, y_train)
prediction = adaboost.predict(x_test)
print("Error rate for %d iterations is %.2f%%" % (epoch, adaboost.evaluate(prediction, y_test)))
epoch = 5
weights = np.ones(len(x_train)) / len(x_train)
adaboost = Adaboost(weights, epoch, DecisionRule())
adaboost.train(x_train, y_train)
prediction = adaboost.predict(x_test)
print("Error rate for %d iterations is %.2f%%" % (epoch, adaboost.evaluate(prediction, y_test)))
epoch = 10
weights = np.ones(len(x_train)) / len(x_train)
adaboost = Adaboost(weights, epoch, DecisionRule())
adaboost.train(x_train, y_train)
prediction = adaboost.predict(x_test)
print("Error rate for %d iterations is %.2f%%" % (epoch, adaboost.evaluate(prediction, y_test)))
epoch = 20
weights = np.ones(len(x_train)) / len(x_train)
adaboost = Adaboost(weights, epoch, DecisionRule())
adaboost.train(x_train, y_train)
prediction = adaboost.predict(x_test)
print("Error rate for %d iterations is %.2f%%" % (epoch, adaboost.evaluate(prediction, y_test)))
def main():
data = DataHelper.Data()
x_train, y_train, x_test, y_test, _ = data.loadData(
"hw2_data_2.txt", 20, 700)
# radial kernel
svm_radial = SVM(x_train,
y_train,
x_test,
y_test,
kernel="RADIAL",
gamma_range=np.logspace(-3, 2, 6))
svm_radial.run()
# sigmoid kernel
svm_sigmoid = SVM(x_train,
y_train,
x_test,
y_test,
kernel="SIGMOID",
gamma_range=np.logspace(-3, 2, 6))
svm_sigmoid.run()
# polynomial kernel
svm_poly = SVM(x_train,
y_train,
x_test,
y_test,
kernel="POLYNOMIAL",
degree_range=range(1, 11))
svm_poly.run()
plt.show()
def main():
data = DataHelper.Data()
x_train, y_train, x_test, y_test, attr_list = data.loadData(
"hw2_data_2.txt", 20, 700)
# print(x_train.shape, y_train.shape, x_test.shape, y_test.shape, attr_list.shape)
gb = GradientBoosting(x_train, y_train, x_test, y_test, attr_list)
gb.fit()
def main():
data = DataHelper.Data()
x_train, y_train, x_test, y_test, _ = data.loadData(
"hw2_data_2.txt", 20, 700)
mars = MARS(x_train, y_train, x_test, y_test)
mars.fit()
print("The testing error rate for MARS classifier is: %.4f" %
mars.evaluate())
def main():
data = DataHelper.Data()
x_train, y_train, x_test, y_test, _ = data.loadData(
"hw2_data_1.txt", 2, 70) # load data
weights = np.ones(x_train.shape[1] + 1)
epochs = 50
model = perceptron(weights, learningRate=1, epoch=epochs)
model.train(x_train, y_train)
print("The error rate for perceptron after %i epochs is %.2f %%" %
(epochs, model.evaluate(x_test, y_test)))
