def test_print_progress_3():
lr = LogisticRegression(epochs=100,
eta=0.01,
minibatches=1,
print_progress=3,
random_seed=1)
lr.fit(X, y)
def test_logistic_regression_gd():
t = np.array([0.51, 1.18, 4.40])
lr = LogisticRegression(epochs=100, eta=0.01, learning='gd', random_seed=0)
lr.fit(X, y) # 0, 1 class
np.testing.assert_almost_equal(lr.w_, t, 2)
assert((y == lr.predict(X)).all())
def test_logistic_regression_gd():
t = np.array([0.51, 1.18, 4.40])
lr = LogisticRegression(epochs=100, eta=0.01, learning='gd', random_seed=0)
lr.fit(X, y) # 0, 1 class
np.testing.assert_almost_equal(lr.w_, t, 2)
assert ((y == lr.predict(X)).all())
def test_print_progress_3():
lr = LogisticRegression(epochs=100,
eta=0.01,
minibatches=1,
print_progress=3,
random_seed=1)
lr.fit(X, y)
def test_predict_proba():
lr = LogisticRegression(epochs=100, eta=0.01, minibatches=1, random_seed=1)
lr.fit(X, y)
idx = [0, 48, 99] # sample labels: 0, 0, 1
y_pred = lr.predict_proba(X[idx])
expect = np.array([0.009, 0.012, 0.993])
np.testing.assert_almost_equal(y_pred, expect, 3)
def test_score_function():
lr = LogisticRegression(epochs=100,
eta=0.01,
minibatches=1,
random_seed=1)
lr.fit(X, y)
acc = lr.score(X, y)
assert acc == 1.0, "Acc: %s" % acc
def test_ary_persistency_in_shuffling():
orig = X.copy()
lr = LogisticRegression(eta=0.01,
epochs=100,
minibatches=len(y),
l2_lambda=1.0,
random_seed=1)
lr.fit(X, y)
np.testing.assert_almost_equal(orig, X, 6)
def test_logistic_regression_sgd():
t = np.array([0.50, 1.16, 4.38])
lr = LogisticRegression(epochs=100,
eta=0.01,
minibatches=len(y),
random_seed=0)
lr.fit(X, y) # 0, 1 class
np.testing.assert_almost_equal(lr.w_, t, 2)
assert ((y == lr.predict(X)).all())
def test_logistic_regression_sgd():
t = np.array([0.53, 1.2, 4.4])
lr = LogisticRegression(epochs=100,
eta=0.01,
minibatches=len(y),
random_seed=1)
lr.fit(X, y) # 0, 1 class
np.testing.assert_almost_equal(lr.w_, t, 2)
assert((y == lr.predict(X)).all())
def test_l2_regularization_gd():
lr = LogisticRegression(eta=0.01, epochs=20,
learning='gd', regularization='l2',
lambda_=1.0, random_seed=0)
lr.fit(X, y)
y_pred = lr.predict(X)
expect_weights = np.array([ 0.252, 1.186, 2.296])
np.testing.assert_almost_equal(lr.w_, expect_weights, 3)
acc = sum(y_pred == y) / len(y)
assert(acc == 1.0)
def test_score_function():
t = np.array([0.52, 1.2, 4.4])
lr = LogisticRegression(epochs=100,
eta=0.01,
minibatches=1,
random_seed=1)
lr.fit(X, y) # 0, 1 class
np.testing.assert_almost_equal(lr.w_, t, 2)
acc = lr.score(X, y)
assert acc == 1.0, "Acc: %s" % acc
def test_predict_proba():
lr = LogisticRegression(epochs=100,
eta=0.01,
minibatches=1,
random_seed=1)
lr.fit(X, y)
idx = [0, 48, 99] # sample labels: 0, 0, 1
y_pred = lr.predict_proba(X[idx])
expect = np.array([0.009, 0.012, 0.993])
np.testing.assert_almost_equal(y_pred, expect, 3)
def test_logistic_regression_gd():
w = np.array([[1.2], [4.4]])
b = np.array([0.52])
lr = LogisticRegression(epochs=100, eta=0.01, minibatches=1, random_seed=1)
lr.fit(X, y)
np.testing.assert_almost_equal(lr.w_, w, 2)
np.testing.assert_almost_equal(lr.b_, b, 2)
y_pred = lr.predict(X)
acc = np.sum(y == y_pred, axis=0) / float(X.shape[0])
assert acc == 1.0, "Acc: %s" % acc
def test_logistic_regression_sgd():
w = np.array([[1.18], [4.38]])
lr = LogisticRegression(epochs=100,
eta=0.01,
minibatches=len(y),
random_seed=1)
lr.fit(X, y) # 0, 1 class
np.testing.assert_almost_equal(lr.w_, w, 2)
y_pred = lr.predict(X)
acc = np.sum(y == y_pred, axis=0) / float(X.shape[0])
assert acc == 1.0, "Acc: %s" % acc
def test_logistic_regression_gd():
t = np.array([0.52, 1.2, 4.4])
lr = LogisticRegression(epochs=100,
eta=0.01,
minibatches=1,
random_seed=1)
lr.fit(X, y) # 0, 1 class
np.testing.assert_almost_equal(lr.w_, t, 2)
y_pred = lr.predict(X)
acc = np.sum(y == y_pred, axis=0) / float(X.shape[0])
assert acc == 1.0, "Acc: %s" % acc
def test_l2_regularization_gd():
lr = LogisticRegression(eta=0.01,
epochs=20,
minibatches=1,
l2_lambda=1.0,
regularization='l2',
random_seed=1)
lr.fit(X, y)
y_pred = lr.predict(X)
expect_weights = np.array([0.303, 1.066, 2.329])
np.testing.assert_almost_equal(lr.w_, expect_weights, 3)
acc = sum(y_pred == y) / len(y)
assert(acc == 1.0)
def test_l2_regularization_gd():
lr = LogisticRegression(eta=0.01,
epochs=20,
minibatches=1,
l2_lambda=1.0,
random_seed=1)
lr.fit(X, y)
y_pred = lr.predict(X)
expect_weights = np.array([0.153, 1.055, 2.284])
np.testing.assert_almost_equal(lr.w_, expect_weights, 3)
y_pred = lr.predict(X)
acc = np.sum(y == y_pred, axis=0) / float(X.shape[0])
assert acc == 1.0, "Acc: %s" % acc
def test_l2_regularization_gd():
lr = LogisticRegression(eta=0.01,
epochs=20,
minibatches=1,
l2_lambda=1.0,
regularization='l2',
random_seed=0)
lr.fit(X, y)
y_pred = lr.predict(X)
expect_weights = np.array([0.115, 1.032, 2.272])
np.testing.assert_almost_equal(lr.w_, expect_weights, 3)
acc = sum(y_pred == y) / len(y)
assert (acc == 1.0)
def test_l2_regularization_sgd():
lr = LogisticRegression(eta=0.01,
epochs=20,
learning='sgd',
l2_lambda=1.0,
regularization='l2',
random_seed=0)
lr.fit(X, y)
y_pred = lr.predict(X)
expect_weights = np.array([0.09, 0.232, 0.35])
np.testing.assert_almost_equal(lr.w_, expect_weights, 2)
acc = sum(y_pred == y) / len(y)
assert (acc == 1.0)
def test_l2_regularization_sgd():
lr = LogisticRegression(eta=0.01,
epochs=100,
minibatches=len(y),
l2_lambda=1.0,
regularization='l2',
random_seed=1)
lr.fit(X, y)
y_pred = lr.predict(X)
expect_weights = np.array([-2.73e-04, 2.40e-01, 3.53e-01])
np.testing.assert_almost_equal(lr.w_, expect_weights, 2)
acc = sum(y_pred == y) / float(len(y))
assert(acc == 0.97)
def __init__(self,
eta=0.01,
epochs=50,
l2_lambda=0.0,
minibatches=1,
random_seed=None,
print_progress=0):
epochs = int(epochs)
warnings.filterwarnings(module='mlxtend*',
action='ignore',
category=FutureWarning)
_LogisticRegression.__init__(self, eta, epochs, l2_lambda, minibatches,
random_seed, print_progress)
BaseWrapperClf.__init__(self)
def test_l2_regularization_sgd():
lr = LogisticRegression(eta=0.01,
epochs=100,
minibatches=len(y),
l2_lambda=1.0,
regularization='l2',
random_seed=0)
lr.fit(X, y)
y_pred = lr.predict(X)
expect_weights = np.array([0., 0.24, 0.35])
np.testing.assert_almost_equal(lr.w_, expect_weights, 2)
acc = sum(y_pred == y) / float(len(y))
assert (acc == 0.97)
def test_refit_weights():
w = np.array([[1.2], [4.4]])
b = np.array([0.52])
lr = LogisticRegression(epochs=50,
eta=0.01,
minibatches=1,
random_seed=1)
lr.fit(X, y)
w1 = lr.w_[0][0]
w2 = lr.w_[0][0]
lr.fit(X, y, init_params=False)
assert w1 != lr.w_[0][0]
assert w2 != lr.w_[1][0]
np.testing.assert_almost_equal(lr.w_, w, 2)
np.testing.assert_almost_equal(lr.b_, b, 2)
def test_invalid_labels_2():
y1 = np.where(y == 0, -1, 1)
lr = LogisticRegression(epochs=15, eta=0.01, random_seed=1)
assert_raises(AttributeError,
'y array must not contain negative labels.\nFound [-1 1]',
lr.fit,
X,
y1,
{(-1, 1)})
def test_invalid_labels_1():
y1 = np.where(y == 0, 2, 1)
lr = LogisticRegression(epochs=15, eta=0.01, random_seed=1)
if sys.version_info >= (3, 0):
objtype = '{(0, 1)}'
else:
objtype = 'set([(0, 1)])'
expect = 'Labels not in %s.\nFound (1, 2)' % objtype
assert_raises(AttributeError, expect, lr.fit, X, y1, {(0, 1)})
def test_l2_regularization_sgd():
lr = LogisticRegression(eta=0.01,
epochs=100,
minibatches=len(y),
l2_lambda=1.0,
random_seed=1)
lr.fit(X, y)
y_pred = lr.predict(X)
expect_weights = np.array([[0.24], [0.35]])
np.testing.assert_almost_equal(lr.w_, expect_weights, 2)
y_pred = lr.predict(X)
acc = np.sum(y == y_pred, axis=0) / float(X.shape[0])
assert acc == 0.97, "Acc: %s" % acc
def test_refit_weights():
w = np.array([[1.2], [4.4]])
b = np.array([0.52])
lr = LogisticRegression(epochs=50, eta=0.01, minibatches=1, random_seed=1)
lr.fit(X, y)
w1 = lr.w_[0][0]
w2 = lr.w_[0][0]
lr.fit(X, y, init_params=False)
assert w1 != lr.w_[0][0]
assert w2 != lr.w_[1][0]
np.testing.assert_almost_equal(lr.w_, w, 2)
np.testing.assert_almost_equal(lr.b_, b, 2)
def test_invalid_labels_1():
y1 = np.where(y == 0, 2, 1)
lr = LogisticRegression(epochs=15, eta=0.01, random_seed=1)
assert_raises(AttributeError, 'Labels not in {(0, 1)}.\nFound (1, 2)',
lr.fit, X, y1, {(0, 1)})
from mlxtend.data import iris_data
from mlxtend.evaluate import plot_decision_regions
from mlxtend.classifier import LogisticRegression
import matplotlib.pyplot as plt
# Loading Data
X, y = iris_data()
X = X[:, [0, 3]] # sepal length and petal width
X = X[0:100] # class 0 and class 1
y = y[0:100] # class 0 and class 1
# standardize
X[:,0] = (X[:,0] - X[:,0].mean()) / X[:,0].std()
X[:,1] = (X[:,1] - X[:,1].mean()) / X[:,1].std()
lr = LogisticRegression(eta=0.01, epochs=100, learning='sgd')
lr.fit(X, y)
plot_decision_regions(X, y, clf=lr)
plt.title('Logistic Regression - Stochastic Gradient Descent')
plt.show()
print(lr.w_)
plt.plot(range(len(lr.cost_)), lr.cost_)
plt.xlabel('Iterations')
plt.ylabel('Missclassifications')
plt.show()
import matplotlib.pyplot as plt
X, y = iris_data()
X = X[:, [0, 3]]
X = X[0:100]
y = y[0:100]
X[:, 0] = (X[:, 0] - X[:, 0].mean()) / X[:, 0].std()
X[:, 1] = (X[:, 1] - X[:, 1].mean()) / X[:, 1].std()
lr = LogisticRegression(
eta=0.1,
l2_lambda=0.0,
epochs=500,
#minibatches=1, # 1 for Gradient Descent
#minibatches=len(y), # len(y) for SGD learning
minibatches=5, # 100/5 = 20 -> minibatch-s
random_seed=1,
print_progress=3)
lr.fit(X, y)
y_pred = lr.predict(X)
print('Last 3 Class Labels: %s' % y_pred[-3:])
y_pred = lr.predict_proba(X)
print('Last 3 Class Labels: %s' % y_pred[-3:])
plot_decision_regions(X, y, clf=lr)
plt.title("Logistic regression - gd")
plt.show()
import matplotlib.pyplot as plt
X, y = iris_data()
X = X[:, [0, 3]]
X = X[0:100]
y = y[0:100]
X[:,0] = (X[:,0] - X[:,0].mean()) / X[:,0].std()
X[:,1] = (X[:,1] - X[:,1].mean()) / X[:,1].std()
lr = LogisticRegression(eta = 0.1,
l2_lambda=0.0,
epochs=500,
#minibatches=1, # 1 for Gradient Descent
#minibatches=len(y), # len(y) for SGD learning
minibatches=5, # 100/5 = 20 -> minibatch-s
random_seed=1,
print_progress=3)
lr.fit(X, y)
y_pred = lr.predict(X)
print('Last 3 Class Labels: %s' % y_pred[-3:])
y_pred = lr.predict_proba(X)
print('Last 3 Class Labels: %s' % y_pred[-3:])
plot_decision_regions(X, y, clf=lr)
plt.title("Logistic regression - gd")
plt.show()
test_len = (int) (len - train_len)
Y_test = Y[ train_len : ]
print("len ", len)
print("Y train " , Y_train.size)
print("X train " , (int) (X_train.size/6))
print("Y test " , Y_test.size)
print("X test " , (int) (X_test.size/6))
print(type(Y_test))
lr = LogisticRegression(eta=0.05,
l2_lambda=0.0,
epochs=50,
minibatches=1, # for Gradient Descent
random_seed=1,
print_progress=3)
lr.fit(X_train, Y_train)
print("...")
pre = lr.predict(X_train)
correct = 0
total = 0
i2 = 0
while(i2 < Y_test.size):
if(Y_test[i2] == pre[i2]):
correct += 1
def test_clone():
log = LogisticRegression()
clone(log)
from mlxtend.plotting import plot_decision_regions
from mlxtend.classifier import LogisticRegression
import matplotlib.pyplot as plt
# Loading Data
X, y = iris_data()
X = X[:, [0, 3]] # sepal length and petal width
X = X[0:100] # class 0 and class 1
y = y[0:100] # class 0 and class 1
# standardize
X[:, 0] = (X[:, 0] - X[:, 0].mean()) / X[:, 0].std()
X[:, 1] = (X[:, 1] - X[:, 1].mean()) / X[:, 1].std()
lr = LogisticRegression(
eta=0.1,
l2_lambda=0.0,
epochs=100,
minibatches=1, # for Gradient Descent
random_seed=1,
print_progress=3)
lr.fit(X, y)
plot_decision_regions(X, y, clf=lr)
plt.title('Logistic Regression - Gradient Descent')
plt.show()
plt.plot(range(len(lr.cost_)), lr.cost_)
plt.xlabel('Iterations')
plt.ylabel('Cost')
plt.show()
