def test_train_ovr(self):
for k in range(2, 6):
with self.subTest(k):
X, Y = test_data.separable_circle_data_set(50, k)
w, b = pvml.one_vs_rest_svm_train(X, Y, 0, lr=1e-1, steps=1000)
Yhat = pvml.one_vs_rest_svm_inference(X, w, b)[0]
self.assertListEqual(Y.tolist(), Yhat.tolist())
def test_knn_auto_k(self):
for k in range(1, 5):
with self.subTest(k):
X, Y = test_data.separable_circle_data_set(24, k)
k, _ = pvml.knn_select_k(X, Y, maxk=7)
self.assertGreater(k, 0)
self.assertLess(k, 8)
def test_batch(self):
X, Y = test_data.separable_circle_data_set(10, 2)
net = pvml.CNN([X.shape[1], 2, 2], [1, 1])
X = X.reshape(X.shape[0], 1, 1, X.shape[1])
net.train(X, Y, lr=1e-1, steps=1000, batch=10)
Yhat, P = net.inference(X)
self.assertListEqual(Y.tolist(), Yhat.tolist())
def test_heteroscedastic_gda(self):
for k in range(1, 5):
with self.subTest(k):
X, Y = test_data.separable_circle_data_set(48, k)
means, icovs, priors = pvml.hgda_train(X, Y)
Yhat, scores = pvml.hgda_inference(X, means, icovs, priors)
self.assertListEqual(Y.tolist(), Yhat.tolist())
self.assertListEqual(priors.tolist(), [1.0 / k] * k)
def test_train1(self):
for k in range(1, 6):
with self.subTest(k):
X, Y = test_data.separable_circle_data_set(50, k)
w, b = pvml.multinomial_logreg_train(X, Y, 0, lr=1e-1, steps=1000)
P = pvml.multinomial_logreg_inference(X, w, b)
Yhat = P.argmax(1)
self.assertListEqual(Y.tolist(), Yhat.tolist())
def test_train_no_pruning(self):
tree = pvml.ClassificationTree()
div = ["gini", "entropy", "error"]
for k in range(1, 6):
with self.subTest(k):
X, Y = test_data.separable_circle_data_set(50, k)
tree.train(X, Y, diversity=div[k % 3], pruning_cv=0)
Yhat = tree.inference(X)[0]
self.assertListEqual(Y.tolist(), Yhat.tolist())
def test_train_ovo(self):
for k in range(2, 6):
with self.subTest(k):
X, Y = test_data.separable_circle_data_set(50, k)
a, b = pvml.one_vs_one_ksvm_train(X,
Y,
"rbf",
0.1,
0,
lr=1e-1,
steps=1000)
Yhat = pvml.one_vs_one_ksvm_inference(X, X, a, b, "rbf",
0.1)[0]
self.assertListEqual(Y.tolist(), Yhat.tolist())
def test_train1(self):
X, Y = test_data.separable_circle_data_set(50, 2)
w, b = pvml.svm_train(X, Y, 0, lr=1e-1, steps=1000)
Yhat, P = pvml.svm_inference(X, w, b)
self.assertListEqual(Y.tolist(), Yhat.tolist())
self.assertListEqual(Yhat.tolist(), (P > 0).tolist())
def test_knn_5(self):
for k in range(1, 5):
with self.subTest(k):
X, Y = test_data.separable_circle_data_set(48, k)
Yhat, P = pvml.knn_inference(X, X, Y, k=5)
self.assertListEqual(Y.tolist(), Yhat.tolist())
def test_train1(self):
X, Y = test_data.separable_circle_data_set(50, 2)
alpha, b = pvml.ksvm_train(X, Y, "rbf", 0.1, 0, lr=1e-1, steps=1000)
Yhat, P = pvml.ksvm_inference(X, X, alpha, b, "rbf", 0.1)
self.assertListEqual(Y.tolist(), Yhat.tolist())
self.assertListEqual(Yhat.tolist(), (P > 0).tolist())
def test_train1(self):
X, Y = test_data.separable_circle_data_set(50, 2)
w, b = pvml.perceptron_train(X, Y)
Yhat, Z = pvml.perceptron_inference(X, w, b)
self.assertListEqual(Y.tolist(), Yhat.tolist())