def test_x_parameter_with_dimension_different_from_2_should_raise_exception_when_predicting(
):
X = np.array([[1, 1], [1000, 1000], [2, 2], [980, 1099], [3, 3],
[1100, 1010]])
y = np.array([1, 2, 1, 2, 1, 2])
X_test = np.array([1, 5, 999, 999, 15, 15, 1500, 1500, 50, 50])
knn = KNN()
knn.fit(X, y)
with pytest.raises(InvalidDimension):
knn.predict(X_test)
def test_x_parameter_with_different_number_of_columns_from_training_x_should_raise_exception_when_predicting(
):
X = np.array([[1, 1], [1000, 1000], [2, 2], [980, 1099], [3, 3],
[1100, 1010]])
y = np.array([1, 2, 1, 2, 1, 2])
X_test = np.array([[1, 5, 5], [999, 999, 999], [15, 15, 15],
[1500, 1500, 1500], [50, 50, 50]])
knn = KNN()
knn.fit(X, y)
with pytest.raises(IncompatibleShape):
knn.predict(X_test)
def test_prediction_with_manhattan_distance():
X = np.array([[1, 1], [1000, 1000], [2, 2], [980, 1099], [3, 3],
[1100, 1010]])
y = np.array([1, 2, 1, 2, 1, 2])
X_test = np.array([[1, 5], [999, 999], [15, 15], [1500, 1500], [50, 50]])
y_test = np.array([1, 2, 1, 2, 1])
knn = KNN(metric='manhattan')
knn.fit(X, y)
predicted = knn.predict(X_test)
assert np.all(predicted == y_test)
def test_using_iris_dataset():
iris = datasets.load_iris()
X = iris.data
y = iris.target
X_train, X_test, y_train, y_test = train_test_split(X,
y,
test_size=0.2,
random_state=42)
model = KNN(k=3, metric="euclidean")
model.fit(X_train, y_train)
y_pred = model.predict(X_test)
acc = round(accuracy_score(y_test, y_pred), 2)
assert acc == 1.0
def test_compare_with_sklearn_knn_when_using_euclidean_distance():
rng = np.random.RandomState(42)
metric = 'euclidean'
X, y = make_classification(n_samples=700, random_state=rng)
X_train, X_test, y_train, y_test = train_test_split(X,
y,
test_size=0.3,
random_state=42)
sk_knn = KNeighborsClassifier(n_neighbors=3, metric=metric)
sk_knn.fit(X_train, y_train)
sk_result = sk_knn.predict(X_test)
knn = KNN(k=3, metric=metric)
knn.fit(X_train, y_train)
y_predicted = knn.predict(X_test)
assert np.all(y_predicted == sk_result)
from sklearn import datasets
from sklearn.model_selection import train_test_split
import numpy as np
# Iris Dataset
iris = datasets.load_iris()
X_train, X_temp, y_train, y_temp = \
train_test_split(iris.data, iris.target, test_size=.3)
X_validation, X_test, y_validation, y_test = \
train_test_split(X_temp, y_temp, test_size=.3)
# Initialize KNN algorithm
neighbor = KNN()
# Feed dataset into KNN
neighbor.fit(X_train,y_train)
# Evaluate prediction of KNN
print(neighbor.predict(X_train))
print(neighbor.score(X_train, y_train))
print(neighbor.score(X_test, y_test))
#####################################################################
# Multi-class classification Toy example
X = np.array([[-1, -1], [-2, -1], [-3, -2], [1, 1], [2, 1], [8, 8]])
y = np.array([1,1,1,0,0,2])
neighbor.fit(X,y)
print(neighbor.predict(X))
print(neighbor.score(X,y))