def test_linear_classifier_predict_not_fitted():
"""Test of `LinearClassifier` class with prediction without fit."""
lc = LinearClassifier(learning_rate=0.1, iterations=1, standardize=False)
X = np.array([[0], [1]])
with pytest.raises(NotFitted):
lc.predict(X)
def test_linear_classifier_dataset_inconsistancy():
"""Test of `LinearClassifier` with dataset inconsistancy."""
lc = LinearClassifier(learning_rate=0.1, iterations=1)
X = np.array([[1], [1]])
y = np.array([[1]])
with pytest.raises(InvalidInput):
lc.fit(X, y)
def test_linear_classifier_score_not_fitted():
"""Test of `LinearClassifier` class with score calculation without fit."""
lc = LinearClassifier(learning_rate=0.1, iterations=1, standardize=False)
X = np.array([[0], [1]])
y = np.array([[0], [1]])
with pytest.raises(NotFitted):
lc.score(X, y) == 1
def test_linear_classifier_multiclass():
"""Test of `LinearClassifier` with multiclass."""
lc = LinearClassifier(learning_rate=0.1, iterations=2)
X = np.array([[0], [1], [2]])
y = np.array([[0], [1], [2]])
lc.fit(X, y)
assert lc.score(X, y) == 1
def test_linear_classifier_predict():
"""Test of `LinearClassifier` class with a prediction."""
lc = LinearClassifier(learning_rate=0.1, iterations=1, standardize=True)
X = np.array([[0], [1]])
y = np.array([[0], [1]])
lc.fit(X, y)
assert np.equal(lc.predict(np.array([0])), np.array([0]))
def test_linear_classifier_standardized():
"""Test of `LinearClassifier` class with standardization."""
lc = LinearClassifier(learning_rate=0.1, iterations=1, standardize=True)
X = np.array([[0], [1]])
y = np.array([[0], [1]])
lc.fit(X, y)
assert lc.score(X, y) == 1
def test_linear_classifier_history_disabled():
"""Test of `LinearClassifier` when history disabled."""
lc = LinearClassifier(learning_rate=0.1, iterations=1, history=False)
assert lc.history is None
X = np.array([[0], [1]])
y = np.array([[0], [1]])
lc.fit(X, y)
assert lc.history is None
def test_linear_classifier_history_enabled():
"""Test of `LinearClassifier` when history enabled."""
lc = LinearClassifier(learning_rate=0.1, iterations=1, history=True)
assert lc.history == []
X = np.array([[0], [1]])
y = np.array([[0], [1]])
lc.fit(X, y)
assert lc.history is not None
# Create the input features and target data frame
df = pd.DataFrame(data=cancer["data"], columns=cancer["feature_names"])
df["target"] = pd.Series(cancer.target)
X = df[["mean radius", "mean texture"]]
X = X.values
y = df["target"]
y = y.values.reshape(X.shape[0], 1)
# Normalize the features
standardize = Standardization()
X = standardize(X)
# Train the model
lc = LinearClassifier(optimizer=SGD(iterations=15, history=True), standardize=False)
lc.fit(X, y)
# Plot the results
figure, (a0, a1) = plt.subplots(
1,
2,
num="Logistic Regression",
figsize=(16, 9),
gridspec_kw={"width_ratios": [3, 1]},
)
figure.suptitle("Logistic Regression on Wisconsin breast cancer dataset", fontsize=16)
M = X[df["target"] == 0]
B = X[df["target"] == 1]
if __name__ == "__main__":
# Load the Boston house prices dataset
iris = datasets.load_iris()
# Create the input features and target data frame
df = pd.DataFrame(iris.data, columns=iris.feature_names)
df["target"] = pd.Series(iris.target)
# Select the input features
X = df[iris.feature_names].values
# Select the target
y = df["target"]
y = y.values.reshape(y.shape[0], 1)
X_train, y_train, X_test, y_test = split_dataset(X, y)
# Alchina regression
print("--- Alchina ---")
lr_alchina = LinearClassifier(learning_rate=0.05,
iterations=2000,
standardize=True)
train_and_report(lr_alchina, X_train, y_train, X_test, y_test)
# Scikit-learn regression
print("--- Scikit-learn ---")
lr_skitlearn = LogisticRegression(solver="liblinear", multi_class="ovr")
train_and_report(lr_skitlearn, X_train, target_reshape(y_train), X_test,
target_reshape(y_test))