diagnose {inarmix} | R Documentation |
Provides diagnostics of the mixture model fit.
diagnose(results)
results |
An object of class "inarmix" |
An object of class "diagnose.inarmix"
which is a list containing the
following components
em.converged |
An indication of whether or not the "EM algorithm" converged. |
niter |
The number of iterations for required for the EM algorithm to converge. |
nclasses |
The number of latent classes used in the model fit. |
loglikhood |
The values of the log-likelihood function at each EM iteration. |
ConvMat |
This is a matrix containing several pieces of information. The columns with GEE in the title indicate if the algorithms for updating the parameters within an EM iteration have converged. The column with l[i] just contains the log-likelihood value. The column with the title "||Psi^2||" gives the norm of the "global estimating equation" - this should be near zero if the algorithm converged. |
Nicholas Henderson
XX <- cbind(rep(1,9),c(0:8)/4) colnames(XX) <- c("const","time") coefs <- rbind(c(-.2,0),c(1.2,.3)) alpha <- c(.2,.2) scale <- c(2,2) mix.prop <- c(.8,.2) testdat <- GenerateMixData(200,coefs,alpha,scale,mix.prop,XX) testfit <- inarmix(y~time,nclasses=2,id=subject,data=testdat,maxiter=3) diagnose(testfit)