def test_logreg():
X_train, Y_train, X_test, Y_test = import_census(CENSUS_FILE_PATH)
num_features = X_train.shape[1]
# Add a bias
X_train_b = np.append(X_train, np.ones((len(X_train), 1)), axis=1)
X_test_b = np.append(X_test, np.ones((len(X_test), 1)), axis=1)
# my_x=np.array([[3,4],[5,6],[7,8],[9,10],[22,22],[12,23]])
# my_y=np.array([0,1,2,0,2,1]).reshape(6,)
# #print(my_y)
# my_x = np.append(my_x, np.ones((len(my_x), 1)), axis=1)
# test_model = LogisticRegression(2, 3, 2, CONV_THRESHOLD)
# #print(test_model.predict(my_x))
# test_model.train(my_x, my_y)
## Logistic Regression ###
#print(X_train_b.shape)
#print(Y_train.shape)
model = LogisticRegression(num_features, NUM_CLASSES, BATCH_SIZE,
CONV_THRESHOLD)
#print(model.loss(X_train_b, Y_train))
num_epochs = model.train(X_train_b, Y_train)
acc = model.accuracy(X_test_b, Y_test) * 100
print("Test Accuracy: {:.1f}%".format(acc))
print("Number of Epochs: " + str(num_epochs))
acc = 0
return acc
def test_logreg():
X_train, Y_train, X_test, Y_test = import_census(CENSUS_FILE_PATH)
num_features = X_train.shape[1]
# Add a bias
X_train_b = np.append(X_train, np.ones((len(X_train), 1)), axis=1)
X_test_b = np.append(X_test, np.ones((len(X_test), 1)), axis=1)
### Logistic Regression ###
model = LogisticRegression(num_features, NUM_CLASSES, BATCH_SIZE,
CONV_THRESHOLD)
num_epochs = model.train(X_train_b, Y_train)
acc = model.accuracy(X_test_b, Y_test) * 100
print("Test Accuracy: {:.1f}%".format(acc))
print("Number of Epochs: " + str(num_epochs))
return acc
def test_logreg():
X_train, Y_train, X_test, Y_test = import_mnist(MNIST_TRAIN_INPUTS_PATH,
MNIST_TRAIN_LABELS_PATH,
MNIST_TEST_INPUTS_PATH,
MNIST_TEST_LABELS_PATH)
num_features = X_train.shape[1]
# Add a bias
X_train_b = np.append(X_train, np.ones((len(X_train), 1)), axis=1)
X_test_b = np.append(X_test, np.ones((len(X_test), 1)), axis=1)
### Logistic Regression ###
print('--------- LOGISTIC REGRESSION w/ SGD ---------')
model = LogisticRegression(num_features, MNIST_CLASSES)
model.train(X_train_b, Y_train)
print("Test Accuracy: {:.1f}%".format(
model.accuracy(X_test_b, Y_test) * 100))
