set.up: Sets up the data based on input data and model specifications

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

Description

This function sets up the data and sets constants that are essentially the same for all models. This is used within the main wrapper function ‘ple.lma’, but can also be run independently. If a user wants to run the functions ‘fit.independence’, ‘fit.rasch’, ‘fit.gpcm’, or ‘fit.nominal’, the set up function should be run prior to using these functions to create required input. Such an approach can speed up replication studies because ‘set.up’ would only need to be run once and the response vector (i.e., named ‘y’) in the Master data frame be replaced by a new one.

Usage

set.up(
  inData,
  model.type,
  inTraitAdj = NULL,
  inItemTraitAdj = NULL,
  tol = NULL,
  starting.sv = NULL,
  starting.phi = NULL
)

Arguments

inData	A person x item Data frame with response patterns
model.type	Type of model to be fit
inTraitAdj	Trait x Trait adjacency matrix (NULL for independence)
inItemTraitAdj	Item x Trait adjacency matrix (NULL for independence)
tol	Tolerence for deteriming convergence (default: 1e-06)
starting.sv	Starting category scale values/fixed scores (default: sum equal to zero and sum of squares equal to 1)
starting.phi	optional: Starting phi matrix (default: identity matrix)

Value

PersonByItem inData (rows are response patterns)

TraitByTrait Trait x Trait adjacency matrix

ItemByTrait Item x Trait adjacency matrix

item.by.trait Need for re-scaling phi.mat

starting.sv An item by number of category matrix with starting values for scale values for nominal model and fixed category scores for gpcm and rasch models

ItemNames Names of items in inData and PersonByItem

LambdaName Short list of lambda names needed for item regressions

NuName Short list of nu names names needed for item regressions

LambdaNames Long list of lambdas using in Master data set

NuNames Long list of nu using in Master data set

PhiNames Names of the unique phi parameters

npersons Number of individual or persons in data

nitems Number of items

ncat Number of categories

nless Number of unique lambdas and unique nus

ntraits Number of traits

Maxnphi Number of phis to estimate

Nstack Length of master data set

pq.mat An array used to computed (weighted) rest-scores

Phi.mat A number of traits x number of traits Phi matrix (defual: the identity matrix)

Master Master data set formated for input to to mlogit

tol Tolerence for deteriming convergence

Examples

 data(dass)
 inData <- dass[1:250,c("d1", "d2", "d3", "a1","a2","a3","s1","s2","s3")]

 #--- to set data up for model of independence
 ind.setup <- set.up(inData, model.type="independence")

 #--- for model specification for uni-dimensional models
 inTraitAdj  <- matrix(1, nrow=1, ncol=1)
 inItemTraitAdj <- matrix(1, nrow=9, ncol=1)

 i.setup <- set.up(inData, model.type='independence')
 
 r.setup <- set.up(inData, model.type='rasch', inTraitAdj,
                  inItemTraitAdj)

 g.setup <- set.up(inData, model.type='gpcm', inTraitAdj,
                  inItemTraitAdj)

 n.setup <- set.up(inData, model.type='nominal', inTraitAdj,
                  inItemTraitAdj)

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