fit.independence: Fits the log-linear model of independence

In pleLMA: Pseudo-Likelihood Estimation of Log-Multiplicative Association Models

Description

This function fits by the log-linear model of independence (i.e., only includes marginal effect terms) using pseudo-likelihood estimation. This provides a baseline model with which to compare other models. The independence maximumn of the loglikehood can be used is a measure of no association. The input to the function is only the Master data set and the names of marginal effect terms and items, all of which are created by the 'set.up' function. This function is called from 'ple.lma' or can be run output of wrapper.

Usage

fit.independence(Master, LambdaNames, LambdaName, ItemNames)

Arguments

Master	Master data set from set.up
LambdaNames	Needed to define formula
LambdaName	Used for column names of matrix estimates
ItemNames	Used for row names of number of item by parameter matrix of estimated Lambda parameters

Value

phi.mlogit Parameters estimates and mlpl = logLike output from mnlogit

fstack Formual used in stacked regression

estimates Item by parameter estimates matrix

mlpl.phi Maximum of log pseudo-likelihood from stacked regression

AIC Akaike information criterion for pseudo-likelihood (smaller is better)

BIC Bayesian information criterion for pseudo-likelihood (smaller is better)

Examples

#--- data and set-up
data(dass)
inData <- dass[1:250,c("d1", "d2", "d3", "a1","a2","a3","s1","s2","s3")]
s <- set.up(inData, model.type='independence')

#--- fit independence model
ind <- fit.independence(s$Master, s$LambdaNames, s$LambdaName, s$ItemNames)

pleLMA documentation built on Oct. 6, 2021, 1:08 a.m.