def test_trunc_log_density(self):
mu = 0
sigma = 1
a = -1
b = 1
data = np.array([-0.5, 0, 0.5])
expected_dist = s.truncnorm(loc=mu, scale=sigma, a=a, b=b)
actual_dist = TruncatedNormalDistribution(mu=mu,
sigma=sigma,
lower=a,
upper=b)
expected = expected_dist.logpdf(data)
actual = actual_dist.log_density(data)
assert (actual == expected).all()
def test_estimate_parameters_trunc_norm_neg_mean(self):
mu = -2
sigma = 1
a = -3
b = -1
data = np.random.normal(loc=mu, scale=sigma, size=1000)
data = np.array(list(filter(lambda x: x > a and x < b, data)))
dist = TruncatedNormalDistribution(mu=mu,
sigma=sigma,
lower=a,
upper=b)
actual = dist.estimate_parameters(data, np.ones((data.shape)))
new_mu = dist.mu
new_sigma = dist.sigma
assert isclose(mu, new_mu, abs_tol=0.5)
assert new_sigma > 0
def test_trunc_log_density_shifted(self):
mu = 3
sigma = 2
a = 0
b = 4
data = np.array([0.5, 1, 2, 3, 3.9])
expected_dist = s.truncnorm(loc=mu,
scale=sigma,
a=(a - mu) / sigma,
b=(b - mu) / sigma)
actual_dist = TruncatedNormalDistribution(mu=mu,
sigma=sigma,
lower=a,
upper=b)
expected = expected_dist.logpdf(data)
actual = actual_dist.log_density(data)
for i in range(len(expected)):
assert isclose(actual[i], expected[i], abs_tol=1e-10)
def test_trunc_log_density_shifted_small_sig_outside_range(self):
mu = 3
sigma = 0.5
a = 1
b = 4
data = np.array([-1, 0, 0.5, 5])
expected_dist = s.truncnorm(loc=mu,
scale=sigma,
a=(a - mu) / sigma,
b=(b - mu) / sigma)
actual_dist = TruncatedNormalDistribution(mu=mu,
sigma=sigma,
lower=a,
upper=b)
expected = expected_dist.logpdf(data)
expected = np.where(expected == np.inf, -9999, expected)
expected = np.where(expected == -np.inf, -9999, expected)
actual = actual_dist.log_density(data)
for i in range(len(expected)):
assert isclose(actual[i], expected[i], abs_tol=1e-10)
def test_truncated_p_values_same_range(self):
data1 = np.random.normal(loc=0, scale=1, size=1000)
data2 = np.random.normal(loc=3, scale=1, size=1000)
data1 = np.array(list(filter(lambda x: x > 0 and x < 4, data1)))
data2 = np.array(list(filter(lambda x: x > 0 and x < 4, data2)))
dist1 = TruncatedNormalDistribution(mu=0.5, sigma=1, lower=0, upper=4)
dist2 = TruncatedNormalDistribution(mu=1.5, sigma=1, lower=0, upper=4)
data = np.concatenate((data1, data2))
dist_list = [dist1, dist2]
mixture = Data_Rep(data, dist_list)
dist_weight = len(data2) / len(data)
expected = [1.0, dist_weight, 0.0]
observed = []
for i in [-1, 1.5, 5]:
observed.append(mixture.get_p_value(i))
for i in range(len(expected)):
assert isclose(expected[i], observed[i], abs_tol=0.05)
def generate_data(x):
# Censored 1
# predata1 = np.random.normal(loc=-1, scale=1, size=100000)
# censored_data_1 = np.where(predata1 <= -2, -2, predata1)
# censored_data_1 = np.where(censored_data_1 >= 3, 3, censored_data_1)
# censored_dist_1 = CensoredNormalDistribution(mu=1.0, sigma=1, lower=-2, upper=3)
# censored_dist_1_rand = CensoredNormalDistribution(mu=-1, sigma=1, lower=-2, upper=3)
# Censored 2
# predata2 = np.random.normal(loc=3, scale=1, size=1000)
# censored_data_2 = np.where(predata2 <= -3, -3, predata2)
# censored_data_2 = np.where(censored_data_2 >= 4, 4, censored_data_2)
# censored_dist_2 = CensoredNormalDistribution(mu=1.5, sigma=1, lower=-3, upper=4)
# Censored 2 over same range as Censored 1
# predata2 = np.random.normal(loc=2, scale=1, size=100000)
# censored_data_2 = np.where(predata2 <= -2, -2, predata2)
# censored_data_2 = np.where(censored_data_2 >= 3, 3, censored_data_2)
# censored_dist_2_rand = CensoredNormalDistribution(mu=2, sigma=1, lower=-2, upper=3)
# Truncated 1
predata3 = np.random.normal(loc=1, scale=1, size=100000)
truncated_data_1 = np.array(
list(filter(lambda x: x > 0 and x < 6, predata3)))
truncated_dist_1_rand = TruncatedNormalDistribution(mu=0.5,
sigma=1,
lower=0,
upper=6)
# Truncated 2
# predata4 = np.random.normal(loc=5, scale=1, size=146500)
# truncated_data_2 = np.array(list(filter(lambda x: x > 4 and x < 6, predata4)))
# truncated_dist_2_rand = TruncatedNormalDistribution(mu=3, sigma=1.004, lower=4, upper=6)
# Truncated 2 over same range as Truncated 1
predata4 = np.random.normal(loc=5, scale=1, size=146500)
truncated_data_2 = np.array(
list(filter(lambda x: x > 0 and x < 6, predata4)))
truncated_dist_2_rand = TruncatedNormalDistribution(mu=1.8,
sigma=1,
lower=0,
upper=6)
# Normal 1
# normal_data_1 = np.random.normal(loc=-1, scale=1, size=100000)
# normal_dist_1 = mixem.distribution.NormalDistribution(mu=0.5, sigma=1)
# normal_dist_1_rand = mixem.distribution.NormalDistribution(mu=-3, sigma=3)
# Normal 2
# normal_data_2 = np.random.normal(loc=3, scale=2, size=100000)
# normal_dist_2 = mixem.distribution.NormalDistribution(mu=3.434, sigma=1.964)
# normal_dist_2_rand = mixem.distribution.NormalDistribution(mu=1, sigma=0.5)
# trunced_normal = np.array(list(filter(lambda x: x > 0 and x < 2, normal_data_2)))
# trunced_normal_dist = TruncatedNormalDistribution(mu=3, sigma=2, lower=0, upper=2)
mixture, joint_pdf, data = organize_data(x, truncated_data_1,
truncated_data_2,
truncated_dist_1_rand,
truncated_dist_2_rand)
return mixture, joint_pdf, data