def test_mean_std(self):
# test 1D case
x = np.array([1, 2, 3, 4, 5])
weights = np.array([0.2, 0.1, 2, 0.5, 1])
ret = inputs.mean_std_weighted(x)
self.assertTrue(abs(ret["mean"] - 3.0) < 1e-8)
self.assertTrue(abs(ret["std"] - np.sqrt(2 * (4 + 1) / 5)) < 1e-8)
ret = inputs.mean_std_weighted(x, np.ones(x.shape))
self.assertTrue(abs(ret["mean"] - 3.0) < 1e-8)
self.assertTrue(abs(ret["std"] - np.sqrt(2 * (4 + 1) / 5)) < 1e-8)
ret = inputs.mean_std_weighted(x, weights)
m = np.sum(weights * x) / np.sum(weights)
s = np.sqrt(np.sum((x - m) ** 2 * weights) / np.sum(weights))
self.assertTrue(abs(ret["mean"] - m) < 1e-8)
self.assertTrue(abs(ret["std"] - s) < 1e-8)
# 2D case
x = np.array([[1, 2], [-0.5, 0.0], [3, -0.55]])
weights = np.array([0.5, 2, 1.55])
ret = inputs.mean_std_weighted(x, weights)
sum_weights = np.sum(weights)
m1 = (1.0 * 0.5 + -0.5 * 2 + 3 * 1.55) / sum_weights
m2 = (2.0 * 0.5 + 0.0 * 2 + -0.55 * 1.55) / sum_weights
self.assertTrue(np.allclose(ret["mean"], [m1, m2]))
s1 = np.sqrt(((1.0 - m1) ** 2 * 0.5 + (-0.5 - m1) ** 2 * 2.0 + (3 - m1) ** 2 * 1.55) / sum_weights)
s2 = np.sqrt(((2 - m2) ** 2 * 0.5 + (0.0 - m2) ** 2 * 2.0 + (-0.55 - m2) ** 2 * 1.55) / sum_weights)
self.assertTrue(np.allclose(ret["std"], [s1, s2]))
def test_pearson_weighted(self):
from scipy.stats import pearsonr
x = np.array([1, 2, 3, 4, 5])
y = np.array([1.0, 1.5, -0.5, 3.4, 2.9])
weights = np.array([1, 0, 0.5, 2, 1.5])
r_no_wgt = pearsonr(x, y)[0]
r_no_wgt_test = evaluation.pearsonr_weighted(x, y)
r_ones_wgt = evaluation.pearsonr_weighted(x, y, np.ones(x.shape))
self.assertTrue(abs(r_no_wgt - r_no_wgt_test) < 1e-12)
self.assertTrue(abs(r_no_wgt - r_ones_wgt) < 1e-12)
xm = mean_std_weighted(x, weights)
ym = mean_std_weighted(y, weights)
r_wgt = np.sum((x - xm['mean']) * (y - ym['mean']) * weights) / np.sum(weights)
self.assertTrue((evaluation.pearsonr_weighted(x, y, weights) - r_wgt) < 1e-12)
def test_pearson_weighted(self):
from scipy.stats import pearsonr
x = np.array([1, 2, 3, 4, 5])
y = np.array([1.0, 1.5, -0.5, 3.4, 2.9])
weights = np.array([1, 0, 0.5, 2, 1.5])
r_no_wgt = pearsonr(x, y)[0]
r_no_wgt_test = evaluation.pearsonr_weighted(x, y)
r_ones_wgt = evaluation.pearsonr_weighted(x, y, np.ones(x.shape))
self.assertTrue(abs(r_no_wgt - r_no_wgt_test) < 1e-12)
self.assertTrue(abs(r_no_wgt - r_ones_wgt) < 1e-12)
xm = mean_std_weighted(x, weights)
ym = mean_std_weighted(y, weights)
r_wgt = np.sum(
(x - xm['mean']) * (y - ym['mean']) * weights) / np.sum(weights)
self.assertTrue(
(evaluation.pearsonr_weighted(x, y, weights) - r_wgt) < 1e-12)
def test_mean_std(self):
# test 1D case
x = np.array([1, 2, 3, 4, 5])
weights = np.array([0.2, 0.1, 2,0.5, 1])
ret = inputs.mean_std_weighted(x)
self.assertTrue(abs(ret['mean'] - 3.0) < 1e-8)
self.assertTrue(abs(ret['std'] - np.sqrt(2 * (4 + 1) / 5)) < 1e-8)
ret = inputs.mean_std_weighted(x, np.ones(x.shape))
self.assertTrue(abs(ret['mean'] - 3.0) < 1e-8)
self.assertTrue(abs(ret['std'] - np.sqrt(2 * (4 + 1) / 5)) < 1e-8)
ret = inputs.mean_std_weighted(x, weights)
m = np.sum(weights * x) / np.sum(weights)
s = np.sqrt(np.sum((x - m)**2 * weights) / np.sum(weights))
self.assertTrue(abs(ret['mean'] - m) < 1e-8)
self.assertTrue(abs(ret['std'] - s) < 1e-8)
# 2D case
x = np.array([[1, 2],
[-0.5, 0.0],
[3, -0.55]])
weights = np.array([0.5, 2, 1.55])
ret = inputs.mean_std_weighted(x, weights)
sum_weights = np.sum(weights)
m1 = (1.0 * 0.5 + -0.5 * 2 + 3 * 1.55) / sum_weights
m2 = (2.0 * 0.5 + 0.0 * 2 + -0.55 * 1.55) / sum_weights
self.assertTrue(np.allclose(ret['mean'], [m1, m2]))
s1 = np.sqrt(((1.0 - m1) ** 2 * 0.5 + (-0.5 - m1)**2 * 2.0 + (3 - m1)**2 * 1.55) / sum_weights)
s2 = np.sqrt(((2 - m2) ** 2 * 0.5 + (0.0 - m2)**2 * 2.0 + (-0.55 - m2)**2 * 1.55) / sum_weights)
self.assertTrue(np.allclose(ret['std'], [s1, s2]))