def test_fail_minibatches():
lr = TfSoftmaxRegression(epochs=100,
eta=0.5,
minibatches=13,
random_seed=1)
lr.fit(X, y)
assert ((y == lr.predict(X)).all())
def test_binary_logistic_regression_gd():
t = np.array([[-0.28, 0.95], [-2.23, 2.4]])
lr = TfSoftmaxRegression(epochs=100, eta=0.5, minibatches=1, random_seed=1)
lr.fit(X_bin, y_bin)
np.testing.assert_almost_equal(lr.weights_, t, 2)
assert (y_bin == lr.predict(X_bin)).all()
def test_multi_logistic_regression_gd_acc():
lr = TfSoftmaxRegression(epochs=100,
eta=0.5,
minibatches=1,
random_seed=1)
lr.fit(X, y)
assert((y == lr.predict(X)).all())
def test_fail_minibatches():
lr = TfSoftmaxRegression(epochs=100,
eta=0.5,
minibatches=13,
random_seed=1)
lr.fit(X, y)
assert((y == lr.predict(X)).all())
def test_train_acc():
lr = TfSoftmaxRegression(epochs=3,
eta=0.5,
minibatches=1,
random_seed=1)
lr.fit(X, y)
exp = [0.47, 0.65, 0.67]
np.testing.assert_almost_equal(exp, lr.train_acc_, decimal=2)
def test_score_function():
lr = TfSoftmaxRegression(epochs=100,
eta=0.5,
minibatches=1,
random_seed=1)
lr.fit(X, y)
acc = lr.score(X, y)
assert acc == 1.0, acc
def test_multi_logistic_regression_gd_weights():
t = np.array([[-0.94, -1.05, 2.73],
[-2.17, 2.01, 2.51]])
lr = TfSoftmaxRegression(epochs=100,
eta=0.5,
minibatches=1,
random_seed=1)
lr.fit(X, y)
np.testing.assert_almost_equal(lr.weights_, t, 2)
def test_multi_logistic_probas():
lr = TfSoftmaxRegression(epochs=200,
eta=0.75,
minibatches=1,
random_seed=1)
lr.fit(X, y)
idx = [0, 50, 149] # sample labels: 0, 1, 2
y_pred = lr.predict_proba(X[idx])
exp = np.array([[0.99, 0.01, 0.0], [0.01, 0.89, 0.1], [0.0, 0.02, 0.98]])
np.testing.assert_almost_equal(y_pred, exp, 2)
def test_binary_logistic_regression_sgd():
t = np.array([[0.35, 0.32], [-7.14, 7.3]])
lr = TfSoftmaxRegression(epochs=100,
eta=0.5,
minibatches=len(y_bin),
random_seed=1)
lr.fit(X_bin, y_bin) # 0, 1 class
np.testing.assert_almost_equal(lr.weights_, t, 2)
assert (y_bin == lr.predict(X_bin)).all()
def test_binary_logistic_regression_sgd():
t = np.array([[0.35, 0.32],
[-7.14, 7.3]])
lr = TfSoftmaxRegression(epochs=100,
eta=0.5,
minibatches=len(y_bin),
random_seed=1)
lr.fit(X_bin, y_bin) # 0, 1 class
np.testing.assert_almost_equal(lr.weights_, t, 2)
assert((y_bin == lr.predict(X_bin)).all())
def test_binary_logistic_regression_gd():
t = np.array([[-0.28, 0.95],
[-2.23, 2.4]])
lr = TfSoftmaxRegression(epochs=100,
eta=0.5,
minibatches=1,
random_seed=1)
lr.fit(X_bin, y_bin)
np.testing.assert_almost_equal(lr.weights_, t, 2)
assert((y_bin == lr.predict(X_bin)).all())
def test_multi_logistic_probas():
lr = TfSoftmaxRegression(epochs=200,
eta=0.75,
minibatches=1,
random_seed=1)
lr.fit(X, y)
idx = [0, 50, 149] # sample labels: 0, 1, 2
y_pred = lr.predict_proba(X[idx])
exp = np.array([[0.99, 0.01, 0.0],
[0.01, 0.89, 0.1],
[0.0, 0.02, 0.98]])
np.testing.assert_almost_equal(y_pred, exp, 2)
def test_init_params():
t = np.array([[-0.28, 0.95], [-2.23, 2.4]])
lr = TfSoftmaxRegression(epochs=50, eta=0.5, minibatches=1, random_seed=1)
lr.fit(X_bin, y_bin)
lr.fit(X_bin, y_bin, init_params=False)
np.testing.assert_almost_equal(lr.w_, t, 2)
assert (y_bin == lr.predict(X_bin)).all()
def _clf_softmax(trX,teX,trY,teY):
print "factors",factors(trX.shape[0])
print "enter no of mini batch"
trY=trY.astype(int)
teY=teY.astype(int)
mini_batch=int(input())
clf = TfSoftmaxRegression(eta=0.75,
epochs=100,
print_progress=True,
minibatches=mini_batch,
random_seed=1)
clf.fit(trX, trY)
pred=clf.predict(teX)
print _f_count(teY),"test f count"
pred=pred.astype(np.int32)
print _f_count(pred),"pred f count"
conf_mat=confusion_matrix(teY, pred)
process_cm(conf_mat, to_print=True)
print precision_score(teY,pred),"Precision Score"
print recall_score(teY,pred),"Recall Score"
print roc_auc_score(teY,pred), "ROC_AUC"
def _clf_softmax(trX, teX, trY, teY):
print "factors", factors(trX.shape[0])
print "enter no of mini batch"
trY = trY.astype(int)
teY = teY.astype(int)
mini_batch = int(input())
clf = TfSoftmaxRegression(eta=0.75,
epochs=100,
print_progress=True,
minibatches=mini_batch,
random_seed=1)
clf.fit(trX, trY)
pred = clf.predict(teX)
print _f_count(teY), "test f count"
pred = pred.astype(np.int32)
print _f_count(pred), "pred f count"
conf_mat = confusion_matrix(teY, pred)
process_cm(conf_mat, to_print=True)
print precision_score(teY, pred), "Precision Score"
print recall_score(teY, pred), "Recall Score"
print roc_auc_score(teY, pred), "ROC_AUC"
def test_init_params():
t = np.array([[-0.28, 0.95],
[-2.23, 2.4]])
lr = TfSoftmaxRegression(epochs=50,
eta=0.5,
minibatches=1,
random_seed=1)
lr.fit(X_bin, y_bin)
lr.fit(X_bin, y_bin, init_params=False)
np.testing.assert_almost_equal(lr.w_, t, 2)
assert (y_bin == lr.predict(X_bin)).all()
import matplotlib
matplotlib.use('TKAgg')
from mlxtend.tf_classifier import TfSoftmaxRegression
from mlxtend.data import iris_data
from mlxtend.evaluate import plot_decision_regions
import matplotlib.pyplot as plt
X, y = iris_data()
X = X[:, [0, 3]]
X[:, 0] = (X[:, 0] - X[:, 0].mean()) / X[:, 0].std()
X[:, 1] = (X[:, 1] - X[:, 1].mean()) / X[:, 1].std()
# Gradient Descent
lr = TfSoftmaxRegression(eta=0.75,
epochs=20,
print_progress=True,
minibatches=len(y),
random_seed=1)
lr.fit(X, y)
plt.plot(range(len(lr.cost_)), lr.cost_)
plt.xlabel('Iterations')
plt.ylabel('Cost')
plt.show()
def test_valid_acc():
lr = TfSoftmaxRegression(epochs=3, eta=0.5, minibatches=1, random_seed=1)
lr.fit(X, y, X_valid=X[:100], y_valid=y[:100])
exp = [0.5, 0.5, 0.5]
np.testing.assert_almost_equal(exp, lr.valid_acc_, decimal=2)
def test_train_acc():
lr = TfSoftmaxRegression(epochs=3, eta=0.5, minibatches=1, random_seed=1)
lr.fit(X, y)
exp = [0.47, 0.65, 0.67]
np.testing.assert_almost_equal(exp, lr.train_acc_, decimal=2)
def test_score_function():
lr = TfSoftmaxRegression(epochs=100, eta=0.5, minibatches=1, random_seed=1)
lr.fit(X, y)
acc = lr.score(X, y)
assert acc == 1.0, acc
def test_multi_logistic_regression_gd_acc():
lr = TfSoftmaxRegression(epochs=100, eta=0.5, minibatches=1, random_seed=1)
lr.fit(X, y)
assert (y == lr.predict(X)).all()
with tf.Session(graph=g) as sess:
sess.run(init)
self.init_time_ = time()
for epoch in range(self.epochs):
if self.minibatches > 1:
n_idx = np.random.permutation(n_idx)
minis = np.array_split(n_idx, self.minibatches)
costs = []
for idx in minis:
_, c = sess.run([train, cost], feed_dict={tf_idx: idx})
costs.append(c)
from mlxtend.tf_classifier import TfSoftmaxRegression
lr = TfSoftmaxRegression(eta=0.75,
epochs=1000,
print_progress=True,
minibatches=1,
random_seed=1)
lr.fit(X, y)
plt.plot(range(len(lr.cost_)), lr.cost_)
plt.xlabel('Iterations')
plt.ylabel('Cost')
plt.show()
from mlxtend.evaluate import plot_decision_regions
plot_decision_regions(X, y, clf=lr)
plt.title('Softmax Regression via Gradient Descent in TensorFlow')
plt.show()
def test_multi_logistic_regression_gd_weights():
t = np.array([[-0.94, -1.05, 2.73], [-2.17, 2.01, 2.51]])
lr = TfSoftmaxRegression(epochs=100, eta=0.5, minibatches=1, random_seed=1)
lr.fit(X, y)
np.testing.assert_almost_equal(lr.weights_, t, 2)
import matplotlib
matplotlib.use('TKAgg')
from mlxtend.tf_classifier import TfSoftmaxRegression
from mlxtend.data import iris_data
from mlxtend.evaluate import plot_decision_regions
import matplotlib.pyplot as plt
X, y = iris_data()
X = X[:, [0, 3]]
X[:,0] = (X[:,0] - X[:,0].mean()) / X[:, 0].std()
X[:,1] = (X[:,1] - X[:,1].mean()) / X[:, 1].std()
# Gradient Descent
lr = TfSoftmaxRegression(eta = 0.75,
epochs=20,
print_progress=True,
minibatches=len(y),
random_seed=1)
lr.fit(X, y)
plt.plot(range(len(lr.cost_)), lr.cost_)
plt.xlabel('Iterations')
plt.ylabel('Cost')
plt.show()
