lmestFormula {LMest}

Formulas for LMest functions

Description

Bulding formulas for lmest, lmestCont, lmestMixed, and lmestMc.

Usage

lmestFormula(data,
              response, manifest = NULL,
              LatentInitial = NULL, LatentTransition = NULL,
              AddInterceptManifest = FALSE,
              AddInterceptInitial = TRUE,
              AddInterceptTransition = TRUE, responseStart = TRUE,
              manifestStart = TRUE, LatentInitialStart = TRUE,
              LatentTransitionStart = TRUE)

Arguments

data

a data.frame or a matrix of data

response

a numeric or character vector indicating the column indices or the names for the response variables

manifest

a numeric or character vector indicating the column indices or the names for the covariates affecting the measurement model

LatentInitial

a numeric or character vector indicating the column indices or the names for the covariates affecting the initial probabilities

LatentTransition

a numeric or character vector indicating the column indices or the names for the covariates affecting the transition probabilities

AddInterceptManifest

a logical value indicating whether the intercept is added to the covariates affecting the measurement model

AddInterceptInitial

a logical value indicating whether the intercept is added to covariates affecting the initial probabilities

AddInterceptTransition

a logical value indicating whether the intercept is added to covariates affecting the transition probabilities

responseStart

a logical value indicating whether the response variables names start with response argument

manifestStart

a logical value indicating whether the covariates names start with manifest argument

LatentInitialStart

a logical value indicating whether the covariates names start with LatentInitial argument

LatentTransitionStart

a logical value indicating whether the covariates names start with LatentTransition argument

Details

Generates formulas for responsesFormula and latentFormula to use in lmest, lmestCont, lmestMixed, and lmestMc.

Value

Returns a list with responsesFormula and latentFormula objects.

Author(s)

Francesco Bartolucci, Silvia Pandolfi, Fulvia Pennoni, Alessio Farcomeni, Alessio Serafini

Examples

data(data_SRHS_long)
names(data_SRHS_long)

# Formula with response srhs and covariates for both initail and transition: 
# gender,race,educational,age.

## LM model with covariates on the latent model
# and with intercepts on the initial and transition probabilities

fm <- lmestFormula(data = data_SRHS_long,
                   response = "srhs",
                   LatentInitial = 3:6, LatentTransition = 3:6)
fm

## LM model with covariates on the latent model
# and without intercepts on the initial and transition probabilities

fm <- lmestFormula(data = data_SRHS_long,
                   response = "srhs",
                   LatentInitial = 3:6, LatentTransition = 3:6,
                   AddInterceptInitial = FALSE,AddInterceptTransition = FALSE)
fm

######

data(data_criminal_sim)
str(data_criminal_sim)

# Formula with only the responses from y1 to y10

fm <- lmestFormula(data = data_criminal_sim,response = "y")$responsesFormula
fm

# Formula with only the responses from y1 to y10 and intercept for manifest

fm <- lmestFormula(data = data_criminal_sim,
                   response = "y",AddInterceptManifest = TRUE)$responsesFormula
fm


## LM model for continous responses

data(data_long_cont)
names(data_long_cont)

# Formula with response Y1, Y2, no covariate for manifest,
# X1 covariates for initail and X2 covariate for transition

fm <- lmestFormula(data = data_long_cont,
                   response = c("Y"),
                   LatentInitial = "X",
                   LatentTransition = "X2")
fm

## Wrong model specification since two variable start with X.
# Check the starts arguments. 

# For the right model:

fm <- lmestFormula(data = data_long_cont,
                   response = c("Y"),
                   LatentInitial = "X1",LatentTransition = "X2")
fm

## or

fm <- lmestFormula(data = data_long_cont,
                   response = c("Y"),
                   LatentInitial = 6,LatentTransition = "X2",
                   LatentInitialStart = FALSE)
fm

## Not run: 

data(data_criminal_sim)
data_criminal_sim <- data.frame(data_criminal_sim)

# Mixed LM model for females

responsesFormula <- lmestFormula(data = data_criminal_sim,
                                 response = "y")$responsesFormula

out <- lmest(responsesFormula = responsesFormula,
             index = c("id","time"),
             data = data_criminal_sim,
             k = 2)

## End(Not run)