stepLCplfm: Latent class probabilistic latent feature analysis of...

In plfm: Probabilistic Latent Feature Analysis

Latent class probabilistic latent feature analysis of three-way three-mode binary data

Description

The function stepLCplfm subsequently applies the LCplfm function to fit disjunctive, conjunctive or additive models with minF up to maxF latent features and minT to maxT latent classes. The results of the estimated models are stored in a list with F X T components.

Usage

	stepLCplfm(minF=1,maxF=3,minT=1,maxT=3,
                   data,maprule="disj",M=5,emcrit1=1e-3,emcrit2=1e-8,
                   model=1,delta=0.0001,printrun=FALSE,Nbootstrap=2000)

Arguments

minF	Minimum number of latent features included in the model.
maxF	Maximum number of latent features included in the model.
minT	Minimum number of latent classes included in the model.
maxT	Maximum number of latent classes included in the model.
data	A I X J X K data array of binary observations. Observation (i,j,k) (i=1,..,I; j=1,..,J; k=1,..,K) indicates whether object j is associated to attribute k according to rater i.
maprule	Fit disjunctive models (maprule="disj"), conjunctive models (maprule="conj") or additive models (maprule="add") .
M	The number of exploratory runs of the EM algorithm using random starting points for each model.
emcrit1	Convergence criterion to be used for the estimation of candidate models in the exploration step.
emcrit2	Convergence criterion to be used for the estimation of the best model in the final analysis.
model	The type of dependency and heterogeneity assumption included in the model. model=1, model=2, model=3 represent models with a constant object-feature classification per person and with, respectively, class-specific object parameters, class-specific attribute parameters, and class-specific object- and attribute parameters. model=4, model=5, model=6 represent models with a constant attribute-feature classification per person and with, respectively, class-specific object parameters, class-specific attribute parameters, and class-specific object- and attribute parameters.
delta	The precision used to compute standard errors of the model parameters with the method of finite differences.
printrun	printrun=TRUE prints the analysis type (disjunctive or conjunctive), the number of features (F), the number of latent classes (T) and the number of the run to the output screen, whereas printrun=FALSE suppresses the printing.
Nbootstrap	Number of bootstrap iterations to be used for simulating the reference distribution of odds-ratio dependency measures.

Details

The results of subsequent LCplfm analyses are stored in a matrix of lists with (maxF-minF+1,maxT-minT+1) components.

Author(s)

Michel Meulders

Examples

## Not run: 
# example 1: analysis on determinants of anger-related behavior

# load anger data
data(anger)

# compute 5 runs of disjunctive latent class probabilistic feature models
# with 1 up to 3 features and with 1 up to 2 latent classes
# assume constant situation classification per person 
# and class-specific situation parameters (i.e. model=1) 

anger.lst<-stepLCplfm(minF=1,maxF=3,minT=1,maxT=2,data=anger$data,
                      maprule="disj",M=5,emcrit1=1e-3,emcrit2=1e-8,model=1)


# visualize BIC of fitted models 

par(pty="s")
plot(anger.lst)

# print overview fit measures for all estimated models

anger.lst

# print model with 3 features and 1 latent class

anger.lst[[3,1]]

## End(Not run)

## Not run: 
# example 2:Perceptual analysis of associations between car models and car attributes

# load car data
data(car)


# compute 5 runs of disjunctive models with 4 features and 1 up to 3 latent classes
# assume constant attribute classification per respondent 
# and class-specific car parameters (i.e. model 4)

car.lst<-stepLCplfm(minF=4,maxF=4,minT=1,maxT=3,data=car$data3w,
                      maprule="disj",M=5,emcrit1=1e-3,emcrit2=1e-8,model=4,printrun=TRUE)


# visualize BIC of fitted models
plot(car.lst)

# print overview of fitmeasures for all fitted models
car.lst

## End(Not run)

plfm documentation built on March 30, 2022, 5:08 p.m.