def test_moment(self):
np.testing.assert_almost_equal(
Stats.moment(self.x, max_order=2),
[np.mean(self.x), np.std(self.x)])
self.assertAlmostEqual(Stats.moment(self.x, order=1), np.mean(self.x))
self.assertAlmostEqual(Stats.moment(self.x), np.mean(self.x))
def test_power_mean(self):
p1 = Stats.power_mean(self.x, self.weights, p=0)
self.assertAlmostEqual(p1, Stats.geometric_mean(self.x, self.weights))
p2 = Stats.power_mean(self.x, self.weights, p=2)
holder_mean_2 = np.sum(np.array(self.weights) *
np.array(self.x)**2)**(1. / 2)
self.assertAlmostEqual(p2, holder_mean_2)
def test_mean(self):
self.assertAlmostEqual(Stats.mean(self.x), np.mean(self.x))
self.assertAlmostEqual(Stats.mean(self.x, self.weights),
np.average(self.x, weights=self.weights))
self.assertAlmostEqual(Stats.inverse_mean(self.x),
np.mean(1. / np.array(self.x)))
self.assertAlmostEqual(Stats.mean(self.x, self.weights),
Stats.average(self.x, self.weights))
def test_mae(self):
self.assertAlmostEqual(
Stats.mean_absolute_deviation(self.x),
np.mean(np.abs(np.array(self.x) - np.mean(self.x))))
def test_mode(self):
self.assertAlmostEqual(Stats.mode(self.x), 1.5)
self.assertAlmostEqual(Stats.mode(self.x, self.weights), 2)
def test_harmonic_mean(self):
self.assertAlmostEqual(
Stats.harmonic_mean(self.x, self.weights),
np.sum(self.weights) /
np.sum(np.array(self.weights) / np.array(self.x)))
def test_shifted_geometric_mean(self):
self.assertAlmostEqual(
Stats.shifted_geometric_mean([0.12, -0.08, 0.02], shift=1),
0.0167215)
def test_geometric_mean(self):
self.assertAlmostEqual(
Stats.geometric_mean(self.x, self.weights),
np.prod(np.power(self.x,
self.weights))**(1. / np.sum(self.weights)))
def test_kurtosis_skew(self):
self.assertAlmostEqual(Stats.skewness(self.x), skew(self.x))
self.assertAlmostEqual(Stats.kurtosis(self.x),
kurtosis(self.x, fisher=False))
def test_max_min_range(self):
self.assertEqual(Stats.max(self.x), 5)
self.assertEqual(Stats.min(self.x), 1)
self.assertEqual(Stats.range(self.x), 4)