def test_NNC_imbalanced():
# Make some training data
train = Lab1.generate_data()
x_train = train[['height', 'width']]
y_train = train['label']
# Make some new data to test with
test = Lab1.generate_data(
random_state=43,
proportion=0.3) # switch to class b as the most common
x_test = test[['height', 'width']]
y_test = test['label']
clf = Lab1.OneNearestNeighborClassifier() # Initialize
clf.fit(x_train, y_train) # Learn from the training data
prediction = clf.predict(x_test) # Make a prediction about the test data
accuracy = np.sum(y_test == prediction) / len(y_test)
target_accuracy = 0.65
assert accuracy >= target_accuracy, f'Accuracy {accuracy} != {target_accuracy}'
def test_NNClassifier():
# Make some training data
train = Lab1.generate_data()
x_train = train[['height', 'width']]
y_train = train['label']
# Make some new data to test with
test = Lab1.generate_data(random_state=43) # we want different test data
x_test = test[['height', 'width']]
y_test = test['label']
classifier = Lab1.OneNearestNeighborClassifier() # Initialize
classifier.fit(x_train, y_train) # Learn from the training data
prediction = classifier.predict(
x_test) # Make a prediction about the test data
accuracy = np.sum(y_test == prediction) / len(y_test)
target_accuracy = 0.75
assert accuracy >= target_accuracy, f'Accuracy {accuracy} != {target_accuracy}'