def test_cife(self):
# Given
n_cols = 5
n_rows = 1000
model = LogisticRegression()
# When
mg = MatrixGenerator()
X, y, costs = mg.generate(n_rows=n_rows,
n_basic_cols=n_cols,
noise_sigmas=[0.1, 0.5],
seed=2)
r = 1
beta = 0.5
dvs = FractionVariableSelector()
dvs.fit(data=X,
target_variable=y,
costs=costs,
r=r,
j_criterion_func='cife',
beta=beta)
dvs.score(model=model, scoring_function=roc_auc_score)
dvs.plot_scores(compare_no_cost_method=True,
model=model,
annotate=True)
# Then
self.assertIsInstance(dvs.variables_selected_order, list)
self.assertEqual(len(dvs.variables_selected_order), len(costs))
self.assertAlmostEqual(sum(costs),
sum(dvs.cost_variables_selected_order))
def test_theoretical_output(self):
integer_matrix = np.array([[0, 1, 0], [0, 1, 0], [0, 1, 2], [0, 1, 3],
[1, 1, 5]])
diverse_target = np.array([0, 0, 0, 0, 1])
costs = [1, 1, 1]
r = 1
fvs = FractionVariableSelector()
fvs.fit(data=integer_matrix,
target_variable=diverse_target,
costs=costs,
r=r,
j_criterion_func='mim')
self.assertEqual(fvs.variables_selected_order[0], 2)
def test_pandas_input(self):
integer_matrix = pd.DataFrame(np.random.randint(0, 10, (100, 3)),
columns=['AA', 'BB', 'CC'])
diverse_target = np.random.randint(0, 2, (100))
costs = {'AA': 10, 'BB': 1, 'CC': 1.5}
r = 1
fvs = FractionVariableSelector()
fvs.fit(data=integer_matrix,
target_variable=diverse_target,
costs=costs,
r=r,
j_criterion_func='mim')
self.assertIsInstance(fvs.variables_selected_order, list)
self.assertEqual(len(fvs.variables_selected_order), len(costs))
def test_numpy_input(self):
integer_matrix = np.random.randint(0, 10, (100, 10))
diverse_target = np.random.randint(0, 10, (100))
costs = [1.76, 0.19, -0.36, 0.96, 0.41, 0.17, -0.36, 0.75, 0.79, -1.38]
r = 1
fvs = FractionVariableSelector()
fvs.fit(data=integer_matrix,
target_variable=diverse_target,
costs=costs,
r=r,
j_criterion_func='mim')
self.assertIsInstance(fvs.variables_selected_order, list)
self.assertEqual(len(fvs.variables_selected_order), len(costs))
def test_stop_budget(self):
integer_matrix = pd.DataFrame(np.random.randint(0, 10, (100, 3)),
columns=['AA', 'BB', 'CC'])
diverse_target = pd.Series(np.random.randint(0, 2, (100)))
costs = {'AA': 2, 'BB': 1.1, 'CC': 1.5}
r = 1
fvs = FractionVariableSelector()
fvs.fit(data=integer_matrix,
target_variable=diverse_target,
costs=costs,
r=r,
j_criterion_func='mim',
budget=2,
stop_budget=True)
self.assertGreater(2, sum(fvs.cost_variables_selected_order))
self.assertGreaterEqual(2, len(fvs.variables_selected_order))
def test_score(self):
integer_matrix = np.random.randint(0, 10, (100, 10))
diverse_target = np.random.randint(0, 2, (100))
costs = [1.76, 0.19, -0.36, 0.96, 0.41, 0.17, -0.36, 0.75, 0.79, -1.38]
r = 1
fvs = FractionVariableSelector()
fvs.fit(data=integer_matrix,
target_variable=diverse_target,
costs=costs,
r=r,
j_criterion_func='mim')
model = LogisticRegression()
fvs.score(model, scoring_function=roc_auc_score)
self.assertEqual(len(fvs.total_scores), len(costs))
def test_regard_to_cost_is_better_cife(self):
# Given
n_cols = 3
n_rows = 1000
model = LogisticRegression()
sigmas = [1, 10, 100]
# When
mg = MatrixGenerator()
X, y, costs = mg.generate(n_rows=n_rows,
n_basic_cols=n_cols,
basic_cost=1,
noise_sigmas=sigmas,
seed=42)
r = 0.8
fvs = FractionVariableSelector()
fvs.fit(data=X,
target_variable=y,
costs=costs,
r=r,
j_criterion_func='cife',
beta=0.05)
fvs.score(model=model, scoring_function=roc_auc_score)
fvs.plot_scores(compare_no_cost_method=True, model=model)
def find_nearest_idx(list, value):
array = np.asarray(list)
idx = (np.abs(array - value)).argmin()
return idx
when_better = []
for i in range(fvs.data.shape[1]):
idx_1_no_cost = i
idx_1_cost = find_nearest_idx(
fvs.total_costs, fvs.no_cost_total_costs[idx_1_no_cost])
if fvs.total_scores[idx_1_cost] > fvs.no_cost_total_scores[
idx_1_no_cost]:
when_better.append(True)
else:
when_better.append(False)
# Then
self.assertTrue(sum(when_better) / len(when_better) >= 0.5)