def test_score_correlation_coefficient():
cc = 1
expected_cc = 1
sigma_cc = 0.1
score_cc = ScoreCorrelationCoefficient(cc, sigma_cc, expected_cc)
assert score_cc.p_s_given_cc == pytest.approx(0.9492499653267421)
cc = 0.5
score_cc = ScoreCorrelationCoefficient(cc, sigma_cc, expected_cc)
assert score_cc.p_s_given_cc == pytest.approx(0.19697339347266518)
cc = 0.5
expected_cc = 0.6
sigma_cc = 0.2
score_cc = ScoreCorrelationCoefficient(cc, sigma_cc, expected_cc)
assert score_cc.p_s_given_cc == pytest.approx(0.6178439917879021)
if 0:
import numpy as np
from matplotlib import pyplot as plt
x = np.linspace(-1, 1)
values = [(0.1, 1), (0.1, 0.8), (0.2, 1), (0.2, 0.8)]
fig, axes = plt.subplots(nrows=2, ncols=2)
for ax, (sigma_cc, expected_cc) in zip(axes.flatten(), values):
ax.set_title("E(CC) = %.1f, sigma(cc) = %.1f" %
(expected_cc, sigma_cc))
p_cc_given_s = []
p_cc_given_not_s = []
p_s_given_cc = []
for _ in x:
score_cc = ScoreCorrelationCoefficient(_, sigma_cc,
expected_cc)
p_cc_given_s.append(score_cc.p_cc_given_s)
p_cc_given_not_s.append(score_cc.p_cc_given_not_s)
p_s_given_cc.append(score_cc.p_s_given_cc)
ax.plot(x, p_cc_given_s, label="p(CC;S)")
ax.plot(x, p_cc_given_not_s, label="p(CC;!S)")
ax.plot(x, p_s_given_cc, label="p(S;CC)")
ylim = max(ax.get_ylim()[1] for ax in axes.flatten())
for ax in axes.flatten():
ax.legend()
ax.set_ylim(ax.get_ylim()[0], ylim)
plt.show()
def __init__(self, cc, sigma_cc, cc_true):
"""Initialise a ScoreSymmetryElement object.
Args:
cc (float): the correlation coefficient for this symmetry element
sigma_cc (float): the estimated error in the correlation coefficient
cc_true (float): the expected value of CC if the symmetry element is present,
E(CC; S)
"""
self.cc = cc
self.sigma_cc = sigma_cc
self.z_cc = self.cc / self.sigma_cc
score_cc = ScoreCorrelationCoefficient(self.cc, self.sigma_cc, cc_true)
self.p_cc_given_s = score_cc.p_cc_given_s
self.p_cc_given_not_s = score_cc.p_cc_given_not_s
self.likelihood = score_cc.p_s_given_cc