Ew1z {bbreg}

Ew1z

Description

Auxiliary function required in the Expectation-Maximization algorithm (E-step) and in the calculation of the Fisher information matrix. It represents the conditional expected value E(W_i^s|Z_i), with s = -1; i.e., latent W_i^(-1) given the observed Z_i.

Usage

Ew1z(z, mu, phi)

Arguments

z

response vector with 0 < z_i < 1.

mu

mean parameter (vector having the same size of z).

phi

precision parameter (vector having the same size of z).

Value

Vector of expected values.