def test_y_true_labels_float(self):
test_val = self.empc_test_val
y = [float(l) for l in self.y_true]
yhat = self.y_score
params = self.empc_params
emp = EMPChurn(y_true=y, y_score=yhat, **params)
empc = emp.empc()
self.assertAlmostEqual(empc, test_val)
def test_empc_y_score_method(self):
"""Test EMPC is computed correctly: y_score given as method input."""
test_val = self.empc_test_val
y = self.y_true
yhat = self.y_score
params = self.empc_params
emp = EMPChurn(y_true=y, **params)
empc = emp.empc(y_score=yhat)
self.assertAlmostEqual(empc, test_val)
def test_y_true_labels_str(self):
test_val = self.empc_test_val
y = ["yes" if y_lbl == 1 else "no" for y_lbl in self.y_true]
yhat = self.y_score
params = self.empc_params
params["case_label"] = "yes"
emp = EMPChurn(y_true=y, y_score=yhat, **params)
empc = emp.empc()
self.assertAlmostEqual(empc, test_val)
def test_empc_y_score_instantiation(self):
"""Test EMPC is computed correctly: y_score given at instantiation."""
test_val = self.empc_test_val
y = self.y_true
yhat = self.y_score
params = self.empc_params
emp = EMPChurn(y_true=y, y_score=yhat, **params)
empc = emp.empc()
self.assertAlmostEqual(empc, test_val)
def test_direct_empc_computation(self):
"""Test EMPC is computed correctly: Immediately accessing method."""
test_val = self.empc_test_val
y = self.y_true
yhat = self.y_score
params = self.empc_params
empc1 = EMPChurn(y_true=y, y_score=yhat, **params).empc()
empc2 = EMPChurn(y_true=y, **params).empc(y_score=yhat)
self.assertAlmostEqual(empc1, test_val)
self.assertAlmostEqual(empc2, test_val)
def test_empc_realistic_data(self):
dat = np.loadtxt("../data/predictions.dat")
y_true = dat[:, 0]
y_score = dat[:, 1]
params = {
"alpha": 6,
"beta": 14,
"clv": 200,
"d": 10,
"f": 1,
"case_label": 1,
}
empc = EMPChurn(y_true, y_score, **params).empc()
self.assertAlmostEqual(empc, 14.9946647690104)
def test_proflogit_patsy(self):
yix = 0
X, y = load_data("../data/two-predictors.dat", yix, False)
form = " + ".join([
"standardize(f{})".format(cix) for cix in range(len(X) + 1)
if cix != yix
])
pfl = ProfLogit(rga_kws={
"niter": 5,
"disp": False,
"random_state": 42
}, )
pfl.fit(X, y)
self.assertEqual(pfl.formula, form)
self.assertTrue("Intercept" in pfl.design_info.column_names)
self.assertTrue(pfl.intercept)
y_score = pfl.predict_proba(X)
empc1 = EMPChurn(y, y_score).empc()
empc2 = pfl.score(X, y)
self.assertAlmostEqual(empc1, empc2)
def test_repeated_empc_computation(self):
"""
Test EMPC is computed correctly: Repeated computation with same object.
"""
test_val = 30.3003222629848
y = np.asarray(self.y_true)
yhat = np.array(self.y_score)
params = self.empc_params
obj = EMPChurn(y_true=y, **params)
empc1 = obj.empc(y_score=yhat)
self.assertAlmostEqual(empc1, test_val)
empc2 = obj.empc()
self.assertAlmostEqual(empc2, test_val)
yhat[0] = 0.287577520124614
empc3 = obj.empc(y_score=yhat)
self.assertAlmostEqual(empc3, 30.3001182343737)
yhat[6] = 0.113703411305323
empc4 = obj.empc(y_score=yhat)
self.assertAlmostEqual(empc4, 29.2002764792561)
def test_empc_with_different_parameters_on_realistic_data(self):
dat = np.loadtxt("../data/predictions.dat")
y_true = dat[:, 0]
y_score = dat[:, 1]
tests = [
(
{
"alpha": 2, # changed
"beta": 14,
"clv": 200,
"d": 10,
"f": 1,
},
5.04518077557418,
),
(
{
"alpha": 6,
"beta": 5, # changed
"clv": 200,
"d": 10,
"f": 1,
},
30.2276738879406,
),
(
{
"alpha": 6,
"beta": 14,
"clv": 50, # changed
"d": 10,
"f": 1,
},
1.62729137977619,
),
(
{
"alpha": 6,
"beta": 14,
"clv": 200,
"d": 100, # changed
"f": 1,
},
1.95497116583216,
),
(
{
"alpha": 6,
"beta": 14,
"clv": 200,
"d": 10,
"f": 15, # changed
},
8.47435415117762,
),
(
{
"alpha": 10, # changed
"beta": 7, # changed
"clv": 300, # changed
"d": 50, # changed
"f": 25, # changed
},
23.8505727106226,
),
]
for params, test_val in tests:
empc = EMPChurn(y_true, y_score, **params).empc()
self.assertAlmostEqual(empc, test_val)
def test_raise_assertion_error_unequal_array_lengths(self):
y = self.y_true[1:]
yhat = self.y_score
with self.assertRaises(AssertionError):
EMPChurn(y_true=y, y_score=yhat).empc()
def test_raise_assertion_error_no_y_score_given(self):
y = self.y_true
with self.assertRaises(AssertionError):
EMPChurn(y_true=y).empc()
def test_empc_simple_test07(self):
"""len(np.unique(y_true)) != 2 (all one)"""
y_true = [1, 1]
y_score = [0.25, 0.75]
with self.assertRaises(AssertionError):
EMPChurn(y_true, y_score)
def test_empc_simple_test05(self):
"""Generate QhullError, i.e., ROCCH is equal to diagonal line."""
y_true = [1, 0]
y_score = [0.5, 0.5]
empc = EMPChurn(y_true, y_score).empc()
self.assertAlmostEqual(empc, 22.5007912244511)
def test_empc_simple_test04(self):
"""Perfect prediction (v2)."""
y_true = [1, 0]
y_score = [1, 0]
empc = EMPChurn(y_true, y_score).empc()
self.assertAlmostEqual(empc, 28.0000000000391)
def test_empc_simple_test03(self):
"""Half incorrect prediction."""
y_true = [1, 0]
y_score = [1, 1]
empc = EMPChurn(y_true, y_score).empc()
self.assertAlmostEqual(empc, 22.5007912244511)
