def main(train_path, eval_path, pred_path):
"""Problem 1(b): Logistic regression with Newton's Method.
Args:
train_path: Path to CSV file containing dataset for training.
eval_path: Path to CSV file containing dataset for evaluation.
pred_path: Path to save predictions.
"""
x_train, y_train = util.load_dataset(train_path, add_intercept=True)
x_eval, y_eval = util.load_dataset(eval_path, add_intercept=True)
# *** START CODE HERE ***
# Train a logistic regression classifier
# Plot decision boundary on top of validation set set
# Use np.savetxt to save predictions on eval set to pred_path
initial_theta = np.zeros(x_train.shape[1])
log_reg = LogisticRegression(step_size=0.2,
max_iter=100,
eps=1e-5,
theta_0=initial_theta,
verbose=True)
log_reg.fit(x_train, y_train)
prediction = log_reg.predict(x_eval)
plot_path = pred_path + ".plot.png"
util.plot(x_eval, y_eval, log_reg.theta, plot_path, correction=1.0)
np.savetxt(pred_path, prediction)
def main(train_path, test_path):
x_train, y_train = util.load_dataset(train_path[0])
if len(train_path) == 2:
x_train2, y_train2 = util.load_dataset(train_path[1])
x_train = np.concatenate((x_train, x_train2), axis=0)
y_train = np.concatenate((y_train, y_train2), axis=0)
# Load the data.
x_test, y_test = util.load_dataset(test_path)
# # delete the bert entries
# x_train=x_train[:,-13:]
# x_test=x_test[:,-13:]
# Define the SVM
clf = LogisticRegression().fit(x_train, y_train)
# Predicting
prediction = clf.predict(x_test)
print(classification_report(y_test, prediction))
prediction = clf.predict(x_train)
print(classification_report(y_train, prediction))
def main(lr, train_path, eval_path, pred_path):
"""Problem 3(d): Poisson regression with gradient ascent.
Args:
lr: Learning rate for gradient ascent.
train_path: Path to CSV file containing dataset for training.
eval_path: Path to CSV file containing dataset for evaluation.
pred_path: Path to save predictions.
"""
# Load training set
# x_train, y_train = util.load_dataset(train_path, add_intercept=False)
x_train, y_train = util.load_dataset(train_path, add_intercept=True)
pr = PoissonRegression(max_iter=10000)
pr.step_size = lr
pr.fit(x_train, y_train)
x_eval, y_eval = util.load_dataset(eval_path, add_intercept=True)
y_pred = np.empty_like(y_eval)
for i in range(len(x_eval)):
y_pred[i] = pr.predict(x_eval[i])
# np.savetxt(pred_path, np.column_stack((x_eval, y_pred)), delimiter=',')
np.savetxt(pred_path, y_pred, delimiter=',')
def main(tau, train_path, eval_path):
"""Problem 5(b): Locally weighted regression (LWR)
Args:
tau: Bandwidth parameter for LWR.
train_path: Path to CSV file containing dataset for training.
eval_path: Path to CSV file containing dataset for evaluation.
"""
# Load training set
x_train, y_train = util.load_dataset(train_path, add_intercept=True)
# *** START CODE HERE ***
# Fit a LWR model
model = LocallyWeightedLinearRegression(tau=tau)
model.fit(x_train, y_train)
# Get MSE value on the validation set
x_val, y_val = util.load_dataset(eval_path, add_intercept=True)
y_pred = model.predict(x_val)
print('p5b mse: ', ((y_val - y_pred)**2).mean(axis=0))
# Plot validation predictions on top of training set
plt.figure()
# No need to save predictions
# Plot data
plt.plot(x_train, y_train, 'bx', linewidth=2)
plt.plot(x_val, y_pred, 'ro', linewidth=2)
plt.xlabel('x')
plt.ylabel('y')
plt.savefig('output/p05b.png')
def main(train_path, eval_path, pred_path):
"""Problem 1(b): Logistic regression with Newton's Method.
Args:
train_path: Path to CSV file containing dataset for training.
eval_path: Path to CSV file containing dataset for evaluation.
pred_path: Path to save predictions.
"""
x_train, y_train = util.load_dataset(train_path, add_intercept=True)
# initial guess of parameters
theta_0 = np.zeros(shape=(3, ))
# get the model
model = LogisticRegression(theta_0=theta_0)
model.fit(x_train, y_train)
# predict using the trained model
x_eval, y_eval = util.load_dataset(eval_path, add_intercept=True)
y_pred = model.predict(x_eval)
# Plot decision boundary on top of validation set set
util.plot(x_eval, y_eval, model.theta, 'output/{ds}_log_reg.pdf'.format(ds=eval_path.split('/')[-1]))
# Use np.savetxt to save predictions on eval set to pred_path
np.savetxt(pred_path, y_pred)
def main(train_path, valid_path, save_path):
"""Problem: Gaussian discriminant analysis (GDA)
Args:
train_path: Path to CSV file containing dataset for training.
valid_path: Path to CSV file containing dataset for validation.
save_path: Path to save predicted probabilities using np.savetxt().
"""
# Load dataset
x_train, y_train = util.load_dataset(train_path, add_intercept=False)
# *** START CODE HERE ***
# Train a GDA classifier
clf = GDA()
clf.fit(x_train, y_train)
# Plot decision boundary on validation set
x_test, y_test = util.load_dataset(valid_path, add_intercept=True)
y_test_pred = clf.predict(x_test[:, 1:])
# plot_decision_boundary(x_test, y_test, clf.theta, save_path)
plot(x_test, y_test, clf.theta,
os.path.splitext(save_path)[0] + '_fig.png')
# Use np.savetxt to save predictions on eval set to save_path
np.savetxt(save_path, y_test_pred)
base, ext = os.path.splitext(save_path)
theta_save_path = base + '_theta' + ext
np.savetxt(theta_save_path, clf.theta)
def main(train_path, eval_path, pred_path, k = 0):
"""Problem 1(b): Logistic regression with Newton's Method.
Args:
train_path: Path to CSV file containing dataset for training.
eval_path: Path to CSV file containing dataset for evaluation.
pred_path: Path to save predictions.
"""
x_train, y_train = util.load_dataset(train_path, add_intercept=True)
x_eval, y_eval = util.load_dataset(eval_path, add_intercept = True)
# *** START CODE HERE ***
# Train a logistic regression classifier
# Plot decision boundary on top of validation set set
# Use np.savetxt to save predictions on eval set to pred_path
clf = LogisticRegression()
theta = clf.fit(x_train,y_train)
p = clf.predict(x_eval)
if(k==0):
np.savetxt(pred_path,p,delimiter = ',')
sp = 'output/p01b_plot'
util.plot(x_eval,y_eval,theta,sp)
elif(k==1):
ind = p < 0.5
p[ind] = 0
index = p >= 0.5
p[index] = 1
return theta,p
def main(train_path, valid_path, save_path):
"""Problem: Gaussian discriminant analysis (GDA)
Args:
train_path: Path to CSV file containing dataset for training.
valid_path: Path to CSV file containing dataset for validation.
save_path: Path to save predicted probabilities using np.savetxt().
"""
# Load dataset
x_train, y_train = util.load_dataset(train_path, add_intercept=False)
x_eval, y_eval = util.load_dataset(valid_path, add_intercept=False)
# *** START CODE HERE ***
# Train a GDA classifier
clf = GDA()
clf.fit(x_train, y_train)
preds = clf.predict(x_eval)
# Plot decision boundary on validation set
theta_ = np.insert(clf.theta, 0, clf.theta_zero)
save_path_ = save_path.strip('.txt')
util.plot(x_eval, y_eval, theta_, save_path_)
# Use np.savetxt to save outputs from validation set to save_path
np.savetxt(save_path, preds)
def main(tau, train_path, eval_path):
"""Problem 5(b): Locally weighted regression (LWR)
Args:
tau: Bandwidth parameter for LWR.
train_path: Path to CSV file containing dataset for training.
eval_path: Path to CSV file containing dataset for evaluation.
"""
# Load training set
x_train, y_train = util.load_dataset(train_path, add_intercept=True)
# *** START CODE HERE ***
model = LocallyWeightedLinearRegression(tau=tau)
model.fit(x_train, y_train)
x_eval, y_eval = util.load_dataset(eval_path, add_intercept=True)
y_pred = model.predict(x_eval)
mse = np.mean((y_pred - y_eval)**2)
print(f'MSE={mse}')
plt.figure()
plt.plot(x_train, y_train, 'bx', linewidth=2)
plt.plot(x_eval, y_pred, 'ro', linewidth=2)
plt.xlabel('x')
plt.ylabel('y')
plt.savefig('output/p05b.png')
def main(train_path, eval_path, pred_path):
"""Problem 1(e): Gaussian discriminant analysis (GDA)
Args:
train_path: Path to CSV file containing dataset for training.
eval_path: Path to CSV file containing dataset for evaluation.
pred_path: Path to save predictions.
"""
# Load dataset
x_train, y_train = util.load_dataset(train_path, add_intercept=False)
# get the model
model = GDA()
model.fit(x_train, y_train)
# predict using the trained model
x_eval, y_eval = util.load_dataset(eval_path, add_intercept=False)
y_pred = model.predict(x_eval)
# Plot decision boundary on top of validation set set
theta = list(model.theta)
theta_0 = [model.theta_0]
util.plot(x_eval, y_eval, theta_0 + theta,
'output/{ds}_GDA.pdf'.format(ds=eval_path.split('/')[-1]))
# Use np.savetxt to save predictions on eval set to pred_path
np.savetxt(pred_path, y_pred)
def main(lr, train_path, eval_path, save_path):
"""Problem: Poisson regression with gradient ascent.
Args:
lr: Learning rate for gradient ascent.
train_path: Path to CSV file containing dataset for training.
eval_path: Path to CSV file containing dataset for evaluation.
save_path: Path to save predictions.
"""
# Load training set
x_train, y_train = util.load_dataset(train_path, add_intercept=True)
# *** START CODE HERE ***
# Fit a Poisson Regression model
model = PoissonRegression(step_size=lr)
model.fit(x_train, y_train)
# Run on the validation set, and use np.savetxt to save outputs to save_path
x_val, y_val = util.load_dataset(eval_path, add_intercept=True)
pred_val = model.predict(x_val)
np.savetxt(save_path, pred_val)
# Plot the result
plt.scatter(x=y_val, y=pred_val, label="Predictions")
plt.xlabel("True Count")
plt.ylabel("Predicted Expected Count")
l = np.array([min(y_val), max(y_val)])
plt.plot(l, l, alpha=0.6, color="red", label="45-degree Line")
plt.legend()
image_path = save_path[:-3] + "png"
plt.savefig(image_path)
def main(train_path, eval_path, pred_path):
"""Problem 1(e): Gaussian discriminant analysis (GDA)
Args:
train_path: Path to CSV file containing dataset for training.
eval_path: Path to CSV file containing dataset for evaluation.
pred_path: Path to save predictions.
"""
# Load dataset
x_train, y_train = util.load_dataset(train_path, add_intercept=False)
x_eval, y_eval = util.load_dataset(eval_path, add_intercept=True)
# *** START CODE HERE ***
# Train a GDA classifier
# Plot decision boundary on validation set
# Use np.savetxt to save outputs from validation set to pred_path
gda = GDA(verbose=True)
gda.fit(x_train, y_train)
prediction = gda.predict(x_eval)
plot_path = pred_path + ".plot.png"
util.plot(x_eval, y_eval, gda.theta, plot_path, correction=1.0)
np.savetxt(pred_path, prediction)
def main(tau, train_path, eval_path):
"""Problem 5(b): Locally weighted regression (LWR)
Args:
tau: Bandwidth parameter for LWR.
train_path: Path to CSV file containing dataset for training.
eval_path: Path to CSV file containing dataset for evaluation.
"""
# Load training set
x_train, y_train = util.load_dataset(train_path, add_intercept=True)
# *** START CODE HERE ***
LWR = LocallyWeightedLinearRegression(0.5)
LWR.fit(x_train, y_train)
x_val, y_val = util.load_dataset(eval_path, add_intercept=True)
y_pred = LWR.predict(x_val)
#plot
plt.figure()
plt.plot(x_train[:,1:], y_train, 'bx')
plt.plot(x_val[:,1:], y_pred, 'ro')
plt.show()
mse = ((y_pred - y_val) ** 2).mean()
print(mse)
def main(tau_values, train_path, valid_path, test_path, pred_path):
"""Problem 5(b): Tune the bandwidth paramater tau for LWR.
Args:
tau_values: List of tau values to try.
train_path: Path to CSV file containing training set.
valid_path: Path to CSV file containing validation set.
test_path: Path to CSV file containing test set.
pred_path: Path to save predictions.
"""
# Load training set
x_train, y_train = util.load_dataset(train_path, add_intercept=True)
x_eval, y_eval = util.load_dataset(valid_path, add_intercept=True)
x_test, y_test = util.load_dataset(test_path, add_intercept=True)
# *** START CODE HERE ***
# Search tau_values for the best tau (lowest MSE on the validation set)
MSE_values = []
model_list = []
# Fit a LWR model with the best tau value
for tau in tau_values:
clf = LocallyWeightedLinearRegression(tau)
clf.fit(x_train, y_train)
y_pred = clf.predict(x_eval)
MSE = np.linalg.norm(y_pred - y_eval)**2 / y_eval.shape[0]
print("tau {}, MSE {}".format(tau, MSE))
MSE_values.append(MSE)
model_list.append(clf)
plot_lwr(x_train, y_train, x_eval, y_pred,
"output/tau_{}.png".format(tau))
idx = np.argmin(MSE_values)
best_model, best_tau = model_list[idx], tau_values[idx]
y_test_pred = best_model.predict(x_test)
test_MSE = np.linalg.norm(y_test_pred - y_test)**2 / y_test.shape[0]
print("best tau {}, MSE on the test split {}".format(best_tau, test_MSE))
def load_model_input(args, fold):
logging.info("Loading preprocessed {} dataset".format(fold))
path = get_model_path(args)
X = load_dataset(path / "X_{}.json".format(fold))
Y = load_dataset(path / "Y_{}.json".format(fold))
return X, np.array(Y)
def parallel(self):
X_train, y_train = load_dataset(self.train_set)
X_test, y_test = load_dataset(self.test_set)
rows = X_train.shape[0]
hor_X, hor_y = horizontal_split_data(X_train, y_train, self.part)
lambda_accurancy = np.zeros(self.part)
weights = np.zeros(self.part)
for i in self.number_of_learners:
for j in self.height:
predict_of_test = np.zeros(y_test.shape[0])
adaboost_set = []
for k in hor_X:
lambda_accurancy[k], adaboost_instance = self.AdaBoost_DP(
hor_X[k], hor_y[k], self.privacy_epsilon[k], j, i,
rows)
adaboost_set.append(adaboost_instance)
weights = get_weight(lambda_accurancy)
for k in range(self.part):
predict_of_test += weights[k] * adaboost_set[k].predict(
X_test)
predict_of_test = [
0.0 if predict <= 0.5 else 1.0
for predict in predict_of_test
]
outputFile = open("adaboostdp_output.txt", 'a')
outputFile.write(
f1_score(y_test, predict_of_test, average="micro") + '\n')
outputFile.write(roc_auc_score(y_test, predict_of_test) + '\n')
outputFile.close()
def transforming_stuff(train_path, valid_path, save_path):
x_train, y_train = util.load_dataset(train_path, add_intercept=False)
x_train2 = np.stack([x_train[:, 0], np.log(x_train[:, 1])]).T
# *** START CODE HERE ***
# Train a GDA classifier
clf = GDA()
clf.fit(x_train2, y_train)
# Plot decision boundary on validation set
x_test, y_test = util.load_dataset(valid_path, add_intercept=True)
y_test_pred = clf.predict(x_test[:, 1:])
x, y = x_test, y_test
plt.figure()
plt.plot(x[y == 0, -2], x[y == 0, -1], 'bx', linewidth=2)
plt.plot(x[y == 1, -2], x[y == 1, -1], 'go', linewidth=2)
x1_min, x1_max = x[:, -2].min(), x[:, -2].max()
x_pts = np.arange(x1_min, x1_max, (x1_max - x1_min) / 100)
theta0, theta1, theta2 = clf.theta
y_pts = np.exp((-1 / theta2) * theta1 * x_pts + theta0)
plt.plot(x_pts, y_pts)
plt.xlabel('x1')
plt.ylabel('x2')
if save_path:
plt.savefig(os.path.splitext(save_path)[0] + '_fig.png')
# Use np.savetxt to save predictions on eval set to save_path
np.savetxt(save_path, y_test_pred)
base, ext = os.path.splitext(save_path)
theta_save_path = base + '_theta' + ext
np.savetxt(theta_save_path, clf.theta)
def main(train_path, valid_path, save_path):
"""Problem: Logistic regression with Newton's Method.
Args:
train_path: Path to CSV file containing dataset for training.
valid_path: Path to CSV file containing dataset for validation.
save_path: Path to save predicted probabilities using np.savetxt().
"""
x_train, y_train = util.load_dataset(train_path, add_intercept=False)
x_valid, y_valid = util.load_dataset(valid_path, add_intercept=False)
x_train = x_train[:, 1:]
x_valid = x_valid[:, 1:]
# normalize the data: (skip binary features)
x_train[:, :-1] = (x_train[:, :-1] - np.mean(
x_train[:, :-1], axis=0)) / np.std(x_train[:, :-1], axis=0)
x_valid[:, :-1] = (x_valid[:, :-1] - np.mean(
x_valid[:, :-1], axis=0)) / np.std(x_valid[:, :-1], axis=0)
# add intercept for logistic regression:
x_train = util.add_intercept(x_train)
x_valid = util.add_intercept(x_valid)
clf = logistic.LogisticRegression(step_size=1, max_iter=100000000)
clf.fit(x_train, y_train)
y_pred_prob = clf.predict(x_valid)
y_pred = y_pred_prob.round()
print(classification_report(y_valid, y_pred))
print(confusion_matrix(y_valid, y_pred))
print(np.sum(y_valid))
np.savetxt(save_path, y_pred)
def main(train_path, valid_path, save_path):
"""Problem: Logistic regression with Newton's Method.
Args:
train_path: Path to CSV file containing dataset for training.
valid_path: Path to CSV file containing dataset for validation.
save_path: Path to save predicted probabilities using np.savetxt().
"""
x_train, y_train = util.load_dataset(train_path, add_intercept=True)
# *** START CODE HERE ***
# Train a logistic regression classifier
# Plot decision boundary on top of validation set
# Use np.savetxt to save predictions on eval set to save_path
model = LogisticRegression()
model.fit(x_train, y_train)
x_val, y_val = util.load_dataset(valid_path, add_intercept=True)
util.plot(x_val,
y_val,
model.theta,
save_path=save_path.replace(".txt", "jpg"))
yhat = model.predict(x_val)
np.savetxt(save_path, yhat)
print(f"LogReg acc: {util.compute_accuracy(y_val, yhat)}")
print(f"LogReg log loss: {util.compute_log_loss(y_val, yhat)}")
def main(train_path, test_path):
# load the dataset
# Load the data.
# Load headers
x_train, y_train = util.load_dataset(train_path[0])
if len(train_path) == 2:
x_train2, y_train2 = util.load_dataset(train_path[1])
x_train = np.concatenate((x_train, x_train2), axis=0)
y_train = np.concatenate((y_train, y_train2), axis=0)
x_test, y_test = util.load_dataset(test_path)
h1 = 300
h2 = 70
h3 = 12
h4 = 8
epoch = 100
# acc_train_all, acc_test_all = simple_nn_all(x_train, y_train, x_test, y_test, h1, h2, h3, h4, epoch)
# acc_train_orig, acc_test_orig = simple_nn_orig(x_train, y_train, x_test, y_test, 10, 10, 10, epoch)
acc_train, acc_test = keras_cat_nn(x_train,
y_train,
x_test,
y_test,
h1,
h2,
h3,
h4,
epoch,
feature=False)
def main(tau, train_path, eval_path):
"""Problem 5(b): Locally weighted regression (LWR)
Args:
tau: Bandwidth parameter for LWR.
train_path: Path to CSV file containing dataset for training.
eval_path: Path to CSV file containing dataset for evaluation.
"""
# Load training set
x_train, y_train = util.load_dataset(train_path, add_intercept=True)
# *** START CODE HERE ***
# Fit a LWR model
model = LocallyWeightedLinearRegression(0.5)
model.fit(x_train, y_train)
# Get MSE value on the validation set
x_val, y_val = util.load_dataset(eval_path, add_intercept=True)
y_pred = model.predict(x_val)
mse = ((y_pred - y_val)**2).mean()
print(mse)
# Plot validation predictions on top of training set
# No need to save anything
# Plot data
import matplotlib.pyplot as plt
plt.figure()
plt.plot(x_train, y_train, 'bx')
plt.plot(x_val, y_pred, 'ro')
def main(train_path, valid_path, save_path):
"""Problem: Gaussian discriminant analysis (GDA)
Args:
train_path: Path to CSV file containing dataset for training.
valid_path: Path to CSV file containing dataset for validation.
save_path: Path to save predicted probabilities using np.savetxt().
"""
# Load dataset
x_train, y_train = util.load_dataset(train_path, add_intercept=False)
# *** START CODE HERE ***
# Train a GDA classifier
# Plot decision boundary on validation set
# Use np.savetxt to save outputs from validation set to save_path
model = GDA()
model.fit(x_train, y_train)
x_val, y_val = util.load_dataset(valid_path, add_intercept=False)
util.plot(x_val,
y_val,
model.theta,
save_path=save_path.replace(".txt", "jpg"))
yhat = model.predict(x_val)
np.savetxt(save_path, yhat)
print(f"GDA acc: {util.compute_accuracy(y_val, yhat)}")
print(f"GDA log loss: {util.compute_log_loss(y_val, yhat)}")
def run_trial_lms(n):
"""Problem: Logistic regression with Newton's Method.
Args:
train_path: Path to CSV file containing dataset for training.
valid_path: Path to CSV file containing dataset for validation.
save_path: Path to save predicted probabilities using np.savetxt().
"""
rates = [0.05,0.01, 0.05, 0.1,0.5,1]
train_path = 'ds1_train.csv'
valid_path = 'ds1_valid.csv'
x_train, y_train = util.load_dataset(train_path, add_intercept=True)
# *** START CODE HERE ***
# Train a logistic regression classifier
LR = LogisticRegression(step_size = rates[n])
it = LR.lms_fit(x_train, y_train)
# Plot decision boundary on top of validation set set
x_valid, y_valid = util.load_dataset(valid_path, add_intercept=True)
ac = LR.predict(x_valid, y_valid)
return n,it,ac
def main(train_path, eval_path, pred_path):
"""Problem 1(e): Gaussian discriminant analysis (GDA)
Args:
train_path: Path to CSV file containing dataset for training.
eval_path: Path to CSV file containing dataset for evaluation.
pred_path: Path to save predictions.
"""
# *** START CODE HERE ***
# Train a GDA classifier
# NOTE Drop x0 = 1 convention used in regression examples
# Will need to account for this to write in terms of theta
x_train, y_train = util.load_dataset(train_path, add_intercept=False)
x_eval, y_eval = util.load_dataset(eval_path, add_intercept=False)
model = GDA()
model.fit(x_train, y_train)
predictions = model.predict(x_eval)
np.savetxt(pred_path, predictions)
# Train Logistic regression classifier
x_train, y_train = util.load_dataset(train_path, add_intercept=True)
x_eval, y_eval = util.load_dataset(eval_path, add_intercept=True)
model2 = LogisticRegression()
model2.fit(x_train, y_train)
# Plot decision boundary on validation set
# Compare decision boundary with logistic
thetas = [model.theta, model2.theta]
fig_path = pred_path[:-4] + "_fig.jpg"
colours = ["red", "orange"]
title = "LinearReg (Orange) vs. GDA (Red)"
util.plot_multiple(x_eval, y_eval, thetas, colours, fig_path, title=title)
def main(train_path, valid_path, save_path):
"""Problem: Gaussian discriminant analysis (GDA)
Args:
train_path: Path to CSV file containing dataset for training.
valid_path: Path to CSV file containing dataset for validation.
save_path: Path to save predicted probabilities using np.savetxt().
"""
# Load dataset
x_train, y_train = util.load_dataset(train_path, add_intercept=False)
# *** START CODE HERE ***
# Train a GDA classifier
model = GDA()
# Fit model to the training data. Define theta
model.fit(x_train, y_train)
# Read validation set
x_val, y_val = util.load_dataset(valid_path, add_intercept=True)
# Save predictions to save path
np.savetxt(save_path, model.predict(x_val))
# Plot boundaries
util.plot(x_val, y_val, model.theta, save_path[:-4])
def main(lr, train_path, eval_path, save_path):
"""Problem: Poisson regression with gradient ascent.
Args:
lr: Learning rate for gradient ascent.
train_path: Path to CSV file containing dataset for training.
eval_path: Path to CSV file containing dataset for evaluation.
save_path: Path to save predictions.
"""
# Load training set
x_train, y_train = util.load_dataset(train_path, add_intercept=True)
x_valid, y_valid = util.load_dataset(eval_path, add_intercept=True)
# Train poisson model
reg = PoissonRegression(step_size=lr)
reg.fit(x_train, y_train)
preds = reg.predict(x_valid)
np.savetxt(save_path, preds)
# plot predictions
plt.scatter(y_valid, preds)
plt.xlabel('True count')
plt.ylabel('Predicted count')
plt.axis('equal')
plt.savefig('poisson.jpg')
def main(train_path, valid_path, save_path):
"""Problem: Gaussian discriminant analysis (GDA)
Args:
train_path: Path to CSV file containing dataset for training.
valid_path: Path to CSV file containing dataset for validation.
save_path: Path to save predicted probabilities using np.savetxt().
"""
# Load dataset
x_train, y_train = util.load_dataset(train_path, add_intercept=False)
# *** START CODE HERE ***
# Train a GDA classifier
# Plot decision boundary on validation set
# Use np.savetxt to save outputs from validation set to save_path
x_val, y_val = util.load_dataset(valid_path, add_intercept=False)
###decomment to normalize the training and validation sets to improve the GDA performance:
# x_train = (x_train - np.mean(x_train, axis=0))/np.std(x_train, axis=0)
# x_val = (x_val - np.mean(x_val, axis=0))/np.std(x_val, axis=0)
# x_train = (x_train - np.min(x_train, axis=0))/(np.max(x_train, axis=0) - np.min(x_train, axis=0))
# x_val = (x_val - np.min(x_val, axis=0))/(np.max(x_val, axis=0) - np.min(x_val, axis=0))
clf = GDA()
clf.fit(x_train, y_train)
y_predict = clf.predict(x_val)
np.savetxt(save_path, y_predict)
util.plot(x_val, (y_predict >= 0.5), clf.theta, save_path[:-4]+ "validation_expected")
#plotting the real distribution
util.plot(x_val, y_val, clf.theta, save_path[:-4] + "validation_real")
def main(lr, train_path, eval_path, pred_path):
"""Problem 3(d): Poisson regression with gradient ascent.
Args:
lr: Learning rate for gradient ascent.
train_path: Path to CSV file containing dataset for training.
eval_path: Path to CSV file containing dataset for evaluation.
pred_path: Path to save predictions.
"""
# Load training set
x_train, y_train = util.load_dataset(train_path, add_intercept=False)
x_eval, y_eval = util.load_dataset(train_path, add_intercept=False)
# *** START CODE HERE ***
# Fit a Poisson Regression model
# Run on the validation set, and use np.savetxt to save outputs to pred_path
initial_theta = np.zeros(x_train.shape[1])
#took 2008 iterations to converge using the given learning rate and default epsilon value
poisson_reg = PoissonRegression(step_size=lr, max_iter=10000, theta_0=initial_theta, verbose=True)
poisson_reg.fit(x_train, y_train)
prediction = poisson_reg.predict(x_eval)
np.savetxt(pred_path, prediction)
# comparing prediction and y_eval copied from:
# https://scikit-learn.org/0.16/auto_examples/plot_cv_predict.html
import matplotlib.pyplot as plt
fig,ax = plt.subplots()
ax.scatter(y_eval, prediction)
ax.set_xlabel('Measured')
ax.set_ylabel('Predicted')
fig.savefig(pred_path + ".comparision.png")
def main(lr, train_path, eval_path, save_path, plot_path):
"""Problem: Poisson regression with gradient ascent.
Args:
lr: Learning rate for gradient ascent.
train_path: Path to CSV file containing dataset for training.
eval_path: Path to CSV file containing dataset for evaluation.
save_path: Path to save predictions.
"""
# Load training set
x_train, y_train = util.load_dataset(train_path, add_intercept=True)
# *** START CODE HERE ***
# Fit a Poisson Regression model
# Run on the validation set, and use np.savetxt to save outputs to save_path
clf = PoissonRegression(step_size=lr)
clf.fit(x_train, y_train)
# Evaluation
x_eval, y_eval = util.load_dataset(eval_path, add_intercept=True)
pred_eval = clf.predict(x_eval)
pred_train = clf.predict(x_train)
np.savetxt(save_path, pred_eval)
# Plot
plot_path_train = plot_path.replace(".png", "_train.png")
plot_path_eval = plot_path.replace(".png", "_eval.png")
util.plot_poisson(y_train, pred_train, plot_path_train)
util.plot_poisson(y_eval, pred_eval, plot_path_eval)
def main():
# load dataset
x_train, y_train = util.load_dataset(training_mode=True)
m = x_train.shape[0]
n = x_train.shape[1]
x_train = x_train.reshape(m, n * n)
x_test, y_test = util.load_dataset(training_mode=False)
m = x_test.shape[0]
n = x_test.shape[1]
x_test = x_test.reshape(m, n * n)
x_train = x_train / 255.0
x_test = x_test / 255.0
# build SVM model
svm = SVC(C=5, gamma=0.05)
start_time = dt.datetime.now()
svm.fit(x_train, y_train)
end_time = dt.datetime.now()
elapsed_time = end_time - start_time
print('Elapsed learning {}'.format(str(elapsed_time)))
# predict
expected = y_test
predicted = svm.predict(x_test)
# confusion matrix
cm = metrics.confusion_matrix(expected, predicted)
print("Confusion matrix:\n%s" % cm)
plot_confusion_matrix(cm)
# get accuracy
print("Training Accuracy={}".format(metrics.accuracy_score(y_train, svm.predict(x_train))))
print("Testing Accuracy={}".format(metrics.accuracy_score(y_test, svm.predict(x_test))))
def registry(filename,nf, ptitle, kfstart=2, kfend=5, kstart=1, kend=5): ''' starts the project. For each fold it calculates mean accuracy, standard deviation and plot the corresponding graph. ''' dataset = load_dataset(filename) kf_accuracy = [] for kf in range(kfstart, kfend+1): kf_accuracy.append(get_Allknn_acc_for_kfold(dataset, kf, kstart, kend, nf)) kf_mean_acc = [sum(acclist)/len(acclist) for acclist in kf_accuracy] sd = [numpy.std(acclist) for acclist in kf_accuracy] for kf, acclist in zip(range(kfstart, kfend+1),kf_accuracy): print kf, "fold validation ===> accuracy of", sum(acclist)/len(acclist) # print kf_mean_acc mean_sd = sum(sd)/len(sd) mean_acc = sum(kf_mean_acc)/len(kf_mean_acc) print "Mean accuracy : ", mean_acc print "Mean S.D : ", mean_sd plot_graph(kf_accuracy, kstart, kend,sd, ptitle)
PROJ_GRAD=False # Should we project gradient on tangent space to to the Stiefel Manifold (Orthogonal matrices)?
RETRACT=False # Should we do retraction step?
THRESHOLD=0 #error threshold in which we do the retraction step
GAIN=1 # a multiplicative constant we add to all orthogonal matrices
RETRACT_SKIP=1 # How many Batches to wait before we do retraction
opt_mathods_set=['SGD','ADAM']
OPT_METHOD=opt_mathods_set[0]
algorithm={'ORT_INIT':ORT_INIT,'PROJ_GRAD':PROJ_GRAD,'RETRACT':RETRACT,'THRESHOLD':THRESHOLD,
'GAIN':GAIN,'RETRACT_SKIP':RETRACT_SKIP,'OPT_METHOD':OPT_METHOD}
params={'network':network,'training':training,'algorithm':algorithm}
DO_SAVE=True # should we save results?
save_file_name=get_file_name(params)
#%% Intialize network model
data, vocab, data_ranges = load_dataset(DATASET)
# define a list of parameters to orthogonalize (recurrent connectivities)
param2orthogonlize=[]
# The number of features is number of different letters + 1 unknown letter
FEATURES_NUM=len(vocab)+1
# Construct network
# Input layer
l_in = lasagne.layers.InputLayer(
(BATCH_SIZE, SEQUENCE_LENGTH-1, FEATURES_NUM)) # the input has -1 sequence elength since we through away the last character (it is only predicted - in the output)
layers_to_concat = []
# All recurrent layer
for dd in range(DEPTH):
if ORT_INIT:
W_in_to_hid_init=lasagne.init.Orthogonal(gain=GAIN)