log_full: Function to compute the full cluster-specific likelihood...

In zhenkewu/rewind: Reconstructing Etiology with Binary Decomposition

Description

This function computes the likelihood WITHOUT integrating over the distribution of component specific parameter (e.g., machine usage profiles). This function conditions upon a few model parameters: the true and false positive rates (theta and psi), the Q matrix and p-the prevalence parameter for each machines.

Usage

log_full(Y, eta_star, Q, p, theta, psi)

Arguments

Y	the data for the current cluster (a subset of observations.)
eta_star	A matrix of M columns (of machines). Multivariate binary indicators of presence or absence of protein landmarks (could be a matrix with rows for multiple )
Q	Q-matrix
p	prevalence parameter for each machine; should be a vector of dimension M.
theta	true positive rates
psi	true positive rates

Value

a vector of likelihood values (one per cluster) given other model parameters.

zhenkewu/rewind documentation built on Sept. 9, 2020, 3:40 p.m.