def __init__(self, α: np.ndarray, z: np.ndarray, bor: float):
"""
:param α: sufficient statistics of the posterior Dirichlet density on model/family frequencies
:param z: posterior probabilities for each subject to belong to each model/family
:param bor: Bayesian omnibus risk p(y|H0)/(p(y|H0)+p(y|H1))
"""
self.attribution = z.copy()
self.frequency_mean = dirichlet.mean(α)
self.frequency_var = dirichlet.var(α)
self.exceedance_probability = exceedance_probability(dirichlet(α))
self.protected_exceedance_probability = self.exceedance_probability * (
1 - bor) + bor / len(α) # (7)
def test_frozen_dirichlet():
np.random.seed(2846)
n = np.random.randint(1, 32)
alpha = np.random.uniform(10e-10, 100, n)
d = dirichlet(alpha)
assert_equal(d.var(), dirichlet.var(alpha))
assert_equal(d.mean(), dirichlet.mean(alpha))
assert_equal(d.entropy(), dirichlet.entropy(alpha))
num_tests = 10
for i in range(num_tests):
x = np.random.uniform(10e-10, 100, n)
x /= np.sum(x)
assert_equal(d.pdf(x[:-1]), dirichlet.pdf(x[:-1], alpha))
assert_equal(d.logpdf(x[:-1]), dirichlet.logpdf(x[:-1], alpha))
def splitCargoes(alpha_s, n_customer, N, V, Cat, IT, IT_num, nv,
Container_vehicle, locationtuples, RP, A, nmaps, dist,
historical_routes, historical_routes_tuples):
# Drichlet distribution of plans
multinomial_ed_s = []
for alpha_s_inside in alpha_s:
drichlet_mean_s = dirichlet.mean(alpha_s_inside)
multinomial_ed_s.append(multinomial.rvs(300, drichlet_mean_s, size=1))
# Choosing Cargoes
final_set = []
for cargoewclassiterator in range(0, len(multinomial_ed_s)):
sorted_cargoes = []
for maxfinder in range(max(multinomial_ed_s[cargoewclassiterator][0]),
-1, -1):
if maxfinder > -1:
finded_cargoes = np.where(
multinomial_ed_s[cargoewclassiterator][0] == maxfinder)
sorted_cargoes.append(finded_cargoes[0])
first_set = []
for hh in sorted_cargoes:
for ff in hh:
try:
first_set.append(IT[cargoewclassiterator][ff])
except:
pass
final_set.append([first_set])
# Cargoes for vehicles
set_for_vehicles = []
for vehicle_num in range(0, nv):
init_set = []
for custom_num in range(0, n_customer):
init_set.append(
final_set[custom_num][0]
[vehicle_num *
round(len(final_set[custom_num][0]) / nv):vehicle_num *
round(len(final_set[custom_num][0]) / nv) +
round(len(final_set[custom_num][0]) / nv)])
set_for_vehicles.append(init_set)
return set_for_vehicles, final_set
def findRoutPlans(alpha, n_customer, N, V, Cat, IT, IT_num, nv,
Container_vehicle, locationtuples, RP, A, nmaps, dist,
historical_routes, historical_routes_tuples):
# Drichlet distribution of routes
multinomial_ed = []
for alpha_inside in alpha:
drichlet_mean = dirichlet.mean(alpha_inside)
multinomial_ed.append(multinomial.rvs(300, drichlet_mean, size=RP * 2))
# Finding Edges
init_final_path = []
final_path = []
for eachroute in multinomial_ed:
semi_final_path = []
for routplansiterator in range(0, len(eachroute)):
sorted_edges = []
for maxifinder in range(max(eachroute[routplansiterator]), -1, -1):
finded = np.where(eachroute[routplansiterator] == maxifinder)
sorted_edges.append(finded[0])
first_path = []
for sedge in sorted_edges:
for seg in sedge:
# Checking for duplicate collection points
if len(first_path) > 0:
first_elements = [i for i, j in first_path]
second_elements = [j for i, j in first_path]
if A[seg][0] not in first_elements and A[seg][
1] not in second_elements:
first_path.append(A[seg])
else:
first_path.append(A[seg])
if len(first_path) > n_customer:
semi_final_path.append(first_path[0:(n_customer + 2)])
# Removing invalid routes
numfailed = []
for rpiterator in range(0, len(semi_final_path)):
starting_point = 0
for jkl in semi_final_path[rpiterator]:
starting_point = starting_point + 1
for lkj in semi_final_path[rpiterator][starting_point:]:
if (jkl[1] == lkj[0] and jkl[0] == lkj[1]):
numfailed.append(rpiterator)
clear_paths = [
i for n, i in enumerate(semi_final_path) if n not in numfailed
]
init_final_path.append(clear_paths)
# Choosing different paths for every vehicle between existing rouyte plans
d = 0
final_path = []
for i in range(0, len(init_final_path)):
a = init_final_path[i][d]
final_path.append(a)
for j in range(i + 1, min(i + 2, len(init_final_path))):
p = []
for k in range(0, len(init_final_path[j])):
p.append(len(set(a) & set(init_final_path[j][k])))
d = p.index(min(p))
return final_path
def compute_LPV_from_parameters(alpha_vector):
M = dlt.mean(alpha_vector)
V = dlt.var(alpha_vector)
LPV = M - 1.65 * np.sqrt(V) # 5-percentile
return np.where(LPV < 0, 0, LPV)
# =============================================================================
# Settings
plot_priors = False
n_samples = 10000
# Observed Data
count_obs = OrderedDict({'id1': 87, 'id2': 34, 'id3': 1})
counts = np.array(list(count_obs.values()), dtype=int)
dirichlet_prior = np.ones_like(
counts) # uninformative prior based on pseudo-counts
dirichlet_posterior = dirichlet_prior + counts
prior_samples = get_samples(dirichlet_prior)
posterior_samples = get_samples(dirichlet_posterior)
print('prior means: %s' % (str(dlt.mean(dirichlet_prior))))
PoM = dlt.mean(dirichlet_posterior)
print('posterior means: %s' % (str(PoM)))
PoV = dlt.var(dirichlet_posterior)
print('posterior variances: %s' % (str(PoV)))
print('naive posterior means: %s' % ((counts + 1) / np.sum(counts + 1))
) # expected from value counts plus assumed prior counts
print('Entropy DLT prior:', dlt.entropy(dirichlet_prior))
print('Entropy DLT posterior:', dlt.entropy(dirichlet_posterior))
if plot_priors:
plt.figure(figsize=(9, 6))
for i, label in enumerate(count_obs.keys()):
ax = plt.hist(prior_samples[:, i],
bins=50,
density=True,
def mean(self): # What is the point of this??
return dirichlet.mean(self.alpha)