def test_split_labelled(self):
dataset = DataSet([[1, 2], [3, 4], [5, 6], [7, 8]],
labels=["b", "b", "b", "a"])
first, second = dataset.split(0.5)
self.assertTrue(first.is_labelled())
assert_that(first.get_labels(), equals_series({0: "b", 1: "b"}))
self.assertTrue(second.is_labelled())
assert_that(second.get_labels(), equals_series({2: "b", 3: "a"}))
def test_copy(self):
dataset1 = DataSet([[1, 2], [3, 4]], labels=pd.Series(["a", "b"]))
dataset2 = dataset1.copy()
dataset2.set_column(1, pd.Series([4, 5]))
assert_that(dataset2, equals_dataset([[1, 4], [3, 5]]))
assert_that(dataset2.get_labels(), equals_series({0: "a", 1: "b"}))
assert_that(dataset1, equals_dataset([[1, 2], [3, 4]]))
assert_that(dataset2.get_labels(), equals_series({0: "a", 1: "b"}))
def test_get_feature_value_counts(self):
df = pd.DataFrame([["Jim", 19, 180], ["John", 18, 177],
["Jack", 19, 185]],
columns=["name", "age", "height"])
dataset = DataSet(df)
age_value_counts = dataset.get_feature_value_counts("age")
assert_that(age_value_counts, equals_series({19: 2, 18: 1}))
name_value_counts = dataset.get_feature_value_counts("name")
assert_that(name_value_counts,
equals_series({"Jim": 1, "John": 1, "Jack": 1}))
def test_compute_iteration(self):
dataset = DataSet([[1, 5], [2, 1], [6, 5]])
centroids = [pd.Series([4, 5]), pd.Series([6, 2])]
new_cent, clusters = clustering._compute_iteration(dataset, centroids,
euclidean)
expected_cent = [{0: 3.5, 1: 5}, {0: 2, 1: 1}]
self.assertEqual(len(new_cent), len(expected_cent))
for i, cent in enumerate(new_cent):
assert_that(cent, equals_series(expected_cent[i]))
assert_that(clusters, equals_series({0: 0, 1: 1, 2: 0}))
def test_gradient_descent_3_parameters(self):
"""
Test based on Assignment 1 of the free online
Stanford Machine Learning online course.
"""
dataset = loader.load(self.relative_to_base("datasets/ex1data2.txt"),
has_ids=False,
has_header=False,
has_labels=True,
delimiter=",")
dataset.normalize_features()
dataset.set_column("bias", pd.Series([1] * dataset.num_samples()))
learning_rate = 1.0
iter = 50
initial_theta = pd.Series({0: 0, 1: 0, "bias": 0})
theta = optimize.gradient_descent(dataset,
initial_theta,
learning_rate,
iterations=iter)
assert_that(
theta,
equals_series(
{
0: 110631.050279,
1: -6649.474271,
"bias": 340412.659574
},
places=6))
def test_get_label_value_counts(self):
dataset = DataSet([[1, 2], [3, 4], [5, 6], [7, 8], [9, 10], [11, 12]],
labels=["a", "b", "b", "c", "a", "b"])
expected = {"a": 2, "b": 3, "c": 1}
value_counts = dataset.get_label_value_counts()
assert_that(value_counts, equals_series(expected))
assert_that(value_counts.index, contains("b", "a", "c"))
def test_gradient_descent_2_parameters(self):
"""
Test based on Assignment 1 of the free online
Stanford Machine Learning online course.
For population = 35,000, we predict a profit of 4519.767868
For population = 70,000, we predict a profit of 45342.450129
Final cost: 4.483388
"""
dataset = loader.load(self.relative_to_base("datasets/ex1data1.txt"),
has_ids=False, has_header=False, has_labels=True,
delimiter=",")
dataset.set_column("bias", pd.Series([1] * dataset.num_samples()))
learning_rate = 0.01
iter = 100
initial_theta = pd.Series({0: 0, "bias": 0})
theta = optimize.gradient_descent(dataset, initial_theta,
learning_rate, iterations=iter)
# assert_that(theta.tolist(), contains(-0.576556, 0.859582))
assert_that(theta, equals_series({0: 0.859582,
"bias": -0.576556},
places=6))
def test_gradient_descent_2_parameters(self):
"""
Test based on Assignment 1 of the free online
Stanford Machine Learning online course.
For population = 35,000, we predict a profit of 4519.767868
For population = 70,000, we predict a profit of 45342.450129
Final cost: 4.483388
"""
dataset = loader.load(self.relative_to_base("datasets/ex1data1.txt"),
has_ids=False,
has_header=False,
has_labels=True,
delimiter=",")
dataset.set_column("bias", pd.Series([1] * dataset.num_samples()))
learning_rate = 0.01
iter = 100
initial_theta = pd.Series({0: 0, "bias": 0})
theta = optimize.gradient_descent(dataset,
initial_theta,
learning_rate,
iterations=iter)
# assert_that(theta.tolist(), contains(-0.576556, 0.859582))
assert_that(theta,
equals_series({
0: 0.859582,
"bias": -0.576556
}, places=6))
def test_filter_by_feature_value_with_labels(self):
features = ["name", "hair colour"]
df = pd.DataFrame([["Bill", "brown"], ["Bob", "black"],
["Jim", "brown"]], columns=features)
dataset = DataSet(df, labels=["SENG", "SENG", "CENG"])
filtered = dataset.value_filter("hair colour", "brown")
assert_that(filtered.get_labels(),
equals_series({0: "SENG", 2: "CENG"}))
def test_classify_all(self):
training_set, sample_0 = self.load_car_data()
sample_1 = {"color": "yellow", "type": "sports", "origin": "domestic"}
dataset = DataSet(pd.DataFrame([sample_0, sample_1]),
labels=["no", "yes"])
classifier = NaiveBayes(training_set)
results = classifier.classify_all(dataset)
assert_that(results.get_classifications(), equals_series({0: "no", 1: "no"}))
self.assertEqual(results.compute_accuracy(), 0.5)
def test_get_labelled_rows(self):
dataset = DataSet([[1, 2], [3, 4], [5, 6], [7, 8]],
labels=["a", "a", "b", "b"])
selection = dataset.get_rows([1, 3])
self.assertEqual(selection.num_samples(), 2)
self.assertTrue(selection.is_labelled())
# TODO incorporate labels equals_series into DataSet matcher?
assert_that(selection, equals_dataset([[3, 4], [7, 8]]))
assert_that(selection.get_labels(), equals_series({1: "a", 3: "b"}))
def test_classify_all_weekends(self):
training = load(self.relative_to_base("/datasets/weekends.data"))
classifier = DecisionTree(training)
index = ['weather', 'parents', 'money']
sample_0 = pd.Series(["windy", "no", "rich"], index=index)
sample_1 = pd.Series(["sunny", "yes", "rich"], index=index)
results = classifier.classify_all(
DataSet(pd.DataFrame([sample_0, sample_1])))
assert_that(results.get_classifications(),
equals_series({0: "shopping", 1: "cinema"}))
def test_kmeans_k_3(self):
dataset = DataSet([[3, 13], [5, 13], [2, 11], [4, 11], [6, 11],
[8, 5], [5, 3], [6, 2], [9, 2], [16, 14], [18, 13],
[16, 11], [19, 10]])
preset_centroids = [pd.Series([4, 9]), pd.Series([10, 6]), pd.Series([17, 9])]
clustered = clustering.kmeans(dataset, k=3, centroids=preset_centroids)
assert_that(clustered.get_cluster_assignments(), equals_series({0: 0,
1: 0, 2: 0, 3: 0, 4: 0, 5: 1,
6: 1, 7: 1, 8: 1, 9: 2, 10: 2,
11: 2, 12: 2}))
def test_filter_by_multiple_labels(self):
features = ["name", "hair colour"]
df = pd.DataFrame([["Will", "black"], ["Rob", "blonde"],
["Bill", "brown"], ["Bob", "black"],
["Jim", "brown"]], columns=features)
dataset = DataSet(df, labels=["ELEC", "SENG", "ELEC", "CENG", "SENG"])
filtered = dataset.label_filter(["SENG", "CENG"])
assert_that(filtered, equals_dataset([["Rob", "blonde"],
["Bob", "black"],
["Jim", "brown"]]))
assert_that(filtered.get_labels(),
equals_series({1: "SENG", 3: "CENG", 4: "SENG"}))
def test_gradient_descent_3_parameters(self):
"""
Test based on Assignment 1 of the free online
Stanford Machine Learning online course.
"""
dataset = loader.load(self.relative_to_base("datasets/ex1data2.txt"),
has_ids=False, has_header=False, has_labels=True,
delimiter=",")
dataset.normalize_features()
dataset.set_column("bias", pd.Series([1] * dataset.num_samples()))
learning_rate = 1.0
iter = 50
initial_theta = pd.Series({0: 0, 1: 0, "bias": 0})
theta = optimize.gradient_descent(dataset, initial_theta,
learning_rate, iterations=iter)
assert_that(theta, equals_series({0: 110631.050279,
1: -6649.474271,
"bias": 340412.659574},
places=6))
def test_load_labelled(self):
dataset = load(self.relative_to_base("datasets/3f_ids_header.csv"))
self.assertTrue(dataset.is_labelled())
labels = dataset.get_labels()
assert_that(labels, equals_series({"V01": "c", "V02": "b", "V03": "b",
"V04": "a"}))
def test_get_label_value_counts_no_labels(self):
dataset = DataSet([[1, 2], [3, 4], [5, 6], [7, 8], [9, 10], [11, 12]])
assert_that(dataset.get_label_value_counts(), equals_series({}))
def test_combine_labels(self):
dataset = DataSet([[1, 2], [3, 4], [5, 6]],
labels=pd.Series(["cat", "crow", "pidgeon"]))
dataset.combine_labels(["crow", "pidgeon"], "bird")
labels = dataset.get_labels()
assert_that(labels, equals_series({0: "cat", 1: "bird", 2: "bird"}))