def main():
# load sample data
X, X_label = multivariate_normal.load_data_with_label()
# split all data into train and test set
X_train, X_test, X_label_train, X_label_test = train_test_split(X, X_label)
# compute theta
m, dim = X.shape
initial_theta = np.insert(np.zeros([dim, 1], dtype=np.float32),
0,
1,
axis=0)
theta = fmin_bfgs(compute_cost,
initial_theta,
args=(X_train, X_label_train))
print theta
# plot test data
colors = ['r', 'b']
for i in range(2):
x, y = X_test[X_label_test == i, 0], X_test[X_label_test == i, 1]
plt.scatter(x, y, color=colors[i], marker='x')
# plot decision boundary
# intercept = theta[0]
a = -1.0 * theta[1] / theta[2]
xx = np.linspace(-20, 10)
yy = a * xx - theta[0] / theta[2]
plt.plot(xx, yy, 'k-')
plt.show()
def main():
# load sample data
X, X_label = multivariate_normal.load_data_with_label()
# split all data into train and test set
X_train, X_test, X_label_train, X_label_test = train_test_split(X, X_label)
# compute theta
m, dim = X.shape
initial_theta = np.insert(np.zeros([dim, 1], dtype=np.float32),
0, 1, axis=0)
theta = fmin_bfgs(compute_cost, initial_theta, args=(X_train,
X_label_train))
print theta
# plot test data
colors = ['r', 'b']
for i in range(2):
x, y = X_test[X_label_test == i, 0], X_test[X_label_test == i, 1]
plt.scatter(x, y, color=colors[i], marker='x')
# plot decision boundary
# intercept = theta[0]
a = -1.0 * theta[1] / theta[2]
xx = np.linspace(-20, 10)
yy = a * xx - theta[0] / theta[2]
plt.plot(xx, yy, 'k-')
plt.show()
def main():
X, X_labels = multivariate_normal.load_data_with_label()
X_train, X_test, y_train, y_test = train_test_split(X, X_labels)
clf = SVC()
clf.fit(X_train, y_train)
pred = clf.predict(X_test)
X_labels_uniq = map(np.str, np.unique(X_labels))
print classification_report(y_test, pred,
target_names=X_labels_uniq)
# plot decision boundary with meshgrid
h = 0.1
x_min, x_max = X_train[:, 0].min() - 1, X_train[:, 0].max() + 1
y_min, y_max = X_train[:, 1].min() - 1, X_train[:, 1].max() + 1
xx, yy = np.meshgrid(np.arange(x_min, x_max, h),
np.arange(y_min, y_max, h))
Z = clf.predict(np.c_[xx.ravel(), yy.ravel()])
Z = Z.reshape(xx.shape)
plt.contourf(xx, yy, Z, cmap=plt.cm.Paired, alpha=0.8)
# plot also the training points
plt.scatter(X_train[:, 0], X_train[:, 1], c=y_train, cmap=plt.cm.Paired)
plt.xlim(x_min, x_max)
plt.ylim(y_min, y_max)
plt.show()
def main():
X, X_labels = multivariate_normal.load_data_with_label()
X_train, X_test, X_labels_train, X_labels_test = train_test_split(X,
X_labels)
clf = GaussianNB()
clf.fit(X_train, X_labels_train)
pred = clf.predict(X_test)
X_labels_uniq = map(np.str, np.unique(X_labels))
print classification_report(X_labels_test, pred,
target_names=X_labels_uniq)
def main():
X, X_labels = multivariate_normal.load_data_with_label()
X_train, X_test, X_labels_train, X_labels_test = train_test_split(
X, X_labels)
clf = GaussianNB()
clf.fit(X_train, X_labels_train)
pred = clf.predict(X_test)
X_labels_uniq = map(np.str, np.unique(X_labels))
print classification_report(X_labels_test,
pred,
target_names=X_labels_uniq)
def main():
# load sample data
X, X_labels = multivariate_normal.load_data_with_label()
X_labels[X_labels == 0] = -1
# input data
X_train, X_test, y_train, y_test = train_test_split(X, X_labels)
# training
model = fit(X_train, y_train)
# show boundary
show_boundary(model, X_train, y_train)
def main():
# load sample data
X, X_labels = multivariate_normal.load_data_with_label()
# training
X_train, X_test, y_train, y_test = train_test_split(X, X_labels)
mean, var, class_prior = fit(X_train, y_train)
print 'mean', mean
print 'var', var
model = defaultdict(np.array)
model['mean'] = mean
model['var'] = var
model['class_prior'] = class_prior
# predict
pred = predict(model, X_test, y_test)
# print result
X_labels_uniq = map(np.str, np.unique(X_labels))
print classification_report(y_test, pred, target_names=X_labels_uniq)
def main():
# load sample data
X, X_labels = multivariate_normal.load_data_with_label()
# training
X_train, X_test, y_train, y_test = train_test_split(X, X_labels)
mean, var, class_prior = fit(X_train, y_train)
print "mean", mean
print "var", var
model = defaultdict(np.array)
model["mean"] = mean
model["var"] = var
model["class_prior"] = class_prior
# predict
pred = predict(model, X_test, y_test)
# print result
X_labels_uniq = map(np.str, np.unique(X_labels))
print classification_report(y_test, pred, target_names=X_labels_uniq)
def main():
# load sample data
X, X_labels = multivariate_normal.load_data_with_label()
n_labels = max(X_labels)+1
n_features = len(X[0])
def main():
# load sample data
X, X_labels = multivariate_normal.load_data_with_label()
n_labels = max(X_labels) + 1
n_features = len(X[0])
