gscaLCA {gscaLCA}

Main function of gscaLCA by using fuzzy clustering GSCA

Description

Fitting a component-based LCA by utilizing fuzzy clustering GSCA algorithm.

Usage

gscaLCA(
  dat,
  varnames = NULL,
  ID.var = NULL,
  num.class = 2,
  num.factor = "EACH",
  Boot.num = 20,
  multiple.Core = FALSE,
  covnames = NULL,
  cov.model = NULL,
  multinomial.ref = "MAX"
)

Arguments

dat

Data that you want to fit the gscaLCA function into.

varnames

A character vector. The names of columns to be used in the gscaLCA function.

ID.var

A character element. The name of ID variable. If ID variable is not specified, the gscaLCA function will search an ID variable in the given data. The ID of observations will be automatically generated as a numeric variable if the data set does not include any ID variable. The default is NULL.

num.class

A numeric element. The number of classes to be identified The default is 2.

num.factor

Either "EACH" or "ALLin1"."EACH" specifies the sitatuion that each indicator is assumed to be its phantom latent variable. "ALLin1" indicates that all variables are assumed to be explained by a common latent variable. The default is "EACH".

Boot.num

The number of bootstraps. The standard errors of parameters are computed from the bootstrap within the gscaLCA algorithm. The default is 20.

multiple.Core

A logical element. TRUE enables to use multiple cores for the bootstrap wehn they are available. The default is FASLE.

covnames

A character vector of covariates. The covariates are used when latent class regression (LCR) is fitted.

cov.model

A numeric vector. The indicator function of latent class regression (LCR) that covariates are involved in fitting the fuzzy clustering GSCA. 1 if gscaLCA is for LCR and otherwise 0.

multinomial.ref

A character element. Options of MAX, MIX, FIRST, and LAST are available for setting a reference group. The default is MAX.

Value

A list of the sample size (N), the number of cluster (C), the number of bootstraps (Boot.num/Boot.num.im), the model fit indices (model.fit), the latent class prevalence (LCprevalence), the item response probability (RespProb), the posterior membership & the predicted class membership (membership), and the graphs of item response probability (plot). When it include covariates, the regression results are also provided.

References

Ryoo, J. H., Park, S., & Kim, S. (2019). Categorical latent variable modeling utilizing fuzzy clustering generalized structured component analysis as an alternative to latent class analysis. Behaviormetrika, 47, 291-306. https://doi.org/10.1007/s41237-019-00084-6

Examples

#AddHealth data with 3 clusters with 500 samples
AH.sample= AddHealth[1:500,]
R3 = gscaLCA (dat = AH.sample,
               varnames = names(AddHealth)[2:6],
               ID.var = "AID",
               num.class = 3,
               num.factor = "EACH",
               Boot.num = 0)
summary(R3)
R3$model.fit      # Model fit
R3$LCprevalence   # Latent Class Prevalence
R3$RespProb       # Item Response Probability
head(R3$membership)     # Membership for all observations

# AddHealth data with 3 clusters with 500 samples with two covariates
R3_2C = gscaLCA (dat = AH.sample,
                 varnames = names(AddHealth)[2:6],
                 ID.var = "AID",
                 num.class = 3,
                 num.factor = "EACH",
                 Boot.num = 0,
                 multiple.Core = FALSE,
                 covnames = names(AddHealth)[7:8], # Gender and Edu
                 cov.model = c(1, 0),   # Only Gender varaible is added to the gscaLCR.
                 multinomial.ref = "MAX")

# To print with the results of multinomial regression with hard partitioning of the gscaLCR,
# use the option of "multinomial.hard".
summary(R3_2C, "multinomial.hard")


# AddHealth data with 2 clusters with 20 bootstraps
R2 = gscaLCA(AddHealth,
             varnames = names(AddHealth)[2:6],
             num.class = 2,
             Boot.num = 20,
             multiple.Core = FALSE) # "multiple.Core = TRUE" is recommended.
# TALIS data with 3 clusters with 20 bootstraps and the "ALLin1" option
T3 = gscaLCA(TALIS,
             varnames = names(TALIS)[2:6],
             num.class = 3,
             num.factor = "ALLin1",
             Boot.num = 20,
             multiple.Core = FALSE) # "multiple.Core = TRUE" is recommended.