Description Usage Arguments Details Value Examples
Discrete heterogeneity model is applicable to scenario in which each subject is assigned a latent class, and fixed parameters are used across subjects in the same class. Multiple stages simply mean that more than one responses are allowed in fitting this model. Typically, one response is nested with another, but it does not have to be this case to fit the model.
1 | LatentStage(nclass, ...)
|
nclass |
numbers of class for all subjects, this determines how many groups of different parameters will be obtained |
... |
X1: A covariate matrix of the first stage, similar for X2, X3, ...; y1: A vector of response variables for the first stage, similar for y2, y3, ...; id1: A vector of subject IDs of the first stage, similar for id2, id3, ... |
Arbitrary number of covariates, repsonses and subject ids can be given in the argument list. However, their number must be the same, i.e. equal to the number of stages in this model. Furthermore, dimensions must match, i.e. row(X1) = length(y1) = length(id1).
A list object containing both arguments and results. lambda
is
the estimate of class proportions, which sum up to 1. beta
contains
estimates, SEs and p-values for all linear parameters. posteriorz
lists the probability of each subject belonging to a specific group, the
estimated class id is determined by which class maximizes the probability.
all.loglik
lists log-likelihood in each iteration. y
is a list
object of responses of all stages. id
is a list object of subject ids
of all stages. x
is a list object of covariates of all stages.
AIC
is the AIC for current model. BIC
is the BIC for current
model. These two can be used for class number selection. runtime
is
the whole elapsed time to fit the model.
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 | data(threestage)
attach(threestage)
mod <- LatentStage(5, X1=stage1[, 4:7],
y1=stage1$Y1, id1=stage1$Person,
X2=stage2[, 4:7],
y2=stage2$Y2, id2=stage2$Person,
X3=stage3[, 4:7],
y3=stage3$Y3, id3=stage3$Person)
data(dating)
attach(dating)
nonmiss <- !is.na(wrote)
mod <- LatentStage(3, y1 = browsed, y2 = wrote[nonmiss],
id1 = respid, id2 = respid[nonmiss],
X1 = agedif, X2 = agedif[nonmiss])
|
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.