DP_EM_E_step: DP EM E-step Function

In ericchen12377/mixDP: Mixture Models with Dirichlet Process

Description

E-step to calculate the expectation of log-likelihood in EM algorithm for DP

Usage

DP_EM_E_step(Ymatrix, Xmatrix, t, N, VarMatrix, Beta, K, Pcluster, VP, alpha)

Arguments

Ymatrix	matrix of dependent variables, multiple dimensions eligible
Xmatrix	matrix of independent variables
t	length of trajectory
N	number of trajectories
VarMatrix	covariance matrix
Beta	coefficient matrix
K	number of clusters
Pcluster	weights of clusters
VP	sequence of i.i.d. random variables distributed as Beta(1,alpha), suppose to be infinite for DP, otherwise truncated by K
alpha	hyperparameter with a constant value

Value

list of outputs used for M-step

ericchen12377/mixDP documentation built on June 23, 2021, 12:02 p.m.