def update_beta_j_wrapper(args):
Y_j, alpha_j, beta_prior = args
prop_beta_j = np.empty(alpha_j.shape[0])
prop_beta_j[0] = 1.
for i in range(1, alpha_j.shape[0]):
prop_beta_j[i] = sample_beta_fc(alpha_j[i], Y_j.T[i], beta_prior.a, beta_prior.b)
return prop_beta_j
def update_beta_l_wrapper(args):
""" Wrapper for projgamma.sample_beta_fc
sample_beta_fc assumes a gamma likelihood with gamma prior for the rate parameter.
sampling is done via full conditional (which has form of a gamma).
"""
return sample_beta_fc(*args)
def update_sigma_j_wrapper(args):
zeta_j, Y_j, xi, tau = args
prop_sigma_j = np.empty(zeta_j.shape)
prop_sigma_j[0] = 1.
for i in range(1, prop_sigma_j.shape[0]):
prop_sigma_j[i] = sample_beta_fc(zeta_j[i], Y_j.T[i], xi[i-1], tau[i-1])
return prop_sigma_j
def update_tau_l_wrapper(args):
return sample_beta_fc(*args)
