def test_get_mc1():
constant = np.random.normal()
mu_1 = 1
n = 20
var_epsilon = 1
num_classes = 100
m_c1 = [calculate_va_continuous.get_mc1(construct_class_df(0, 0, n, 0, 0, mu_1, var_epsilon)) \
for i in range(num_classes)]
estimate = np.mean(m_c1)
se = (np.var(m_c1) / num_classes)**(.5)
assert estimate > mu_1 - 2*se
assert estimate < mu_1 + 2*se
def test_get_mc1_precision():
var_epsilon = abs(np.random.random())
std_theta = abs(np.random.random())
n_classes = 100
errors = []
precisions = []
for class_ in range(n_classes):
mu_0 = np.random.random()
mu_1 = np.random.random()
n_students = 20
df = construct_class_df(0, class_, n_students, std_theta, mu_0, mu_1, var_epsilon)
m_c1 = calculate_va_continuous.get_mc1(df)
errors.append(m_c1 - mu_1)
precisions.append(calculate_va_continuous.get_mc1_precision(calculate_va_continuous.collapse(df), var_epsilon))
check_calibration(np.array(errors), np.array(precisions))
def test_get_mc0_precision():
std_theta = 1
var_epsilon = 1
n_students = 20
n_classes = 10
errors = np.zeros(n_classes)
precisions = np.zeros(n_classes)
for class_ in range(n_classes):
mu_1 = np.random.normal()
mu_0 = np.random.normal()
df = construct_class_df(0, 0, n_students, std_theta, mu_0, mu_1, var_epsilon)
collapsed = calculate_va_continuous.collapse(df)
collapsed['m_c1'] = calculate_va_continuous.get_mc1(df)
mc_0 = calculate_va_continuous.get_mc0(collapsed)
errors[class_] = mc_0 - mu_0
precisions[class_] = calculate_va_continuous.get_mc0_precision(std_theta**2, var_epsilon, df['continuous var'].values)
check_calibration(errors, precisions)
