# incompleteSEM Simulation Study
# Seth Temple, sdtemple@uw.edu
# Setup -------------------------------------------------------------------
library(incompleteSEM)
# epidemic model parameters
N <- 200
beta <- 4
gamma <- 1
rates <- c(beta, gamma)
# simulation parameters
simCt <- 0
simSize <- 100
simMin <- 10
simP <- 3/4
simQ <- 1
simPrint <- 10
# storage parameters
simStore <- array(NA, dim = c(simSize,3,3))
simI <- rep(NA, simSize)
# testing -----------------------------------------------------------------
# epi <- rgsem(c(beta, gamma), N)
# epi <- filter_gsem(epi)
#
# likelihood_complete_gsem(rates, epi[,2], epi[,1], N)
# mle_complete_gsem(epi[,2], epi[,1], N)
# pbla_incomplete_gsem(rates, epi[,2], epi[,1], N)
#
# pbla_partial_gsem(rates, epi[,2], N)
#
# epi <- sort_gsem(epi)
#
# epi2 <- impute_gsem(epi, 1/2)
# pbla_incomplete_gsem(rates, epi2[,2], epi2[,1], N)
#
#
# nlm(pbla_incomplete_gsem, c(1,1), r=epi2[,2], i=epi2[,1], N=N)$estimate
# nlm(pbla_partial_gsem, c(1,1), r=epi2[,2], N=N)$estimate
# mle_complete_gsem(epi[,2], epi[,1], N)
# Simulation Study --------------------------------------------------------
while(simCt < simSize){
epi <- rgsem(rates, N)
epi <- filter_gsem(epi)
while(dim(epi)[1] < simMin){
epi <- rgsem(rates, N)
epi <- filter_gsem(epi)
}
simCt <- simCt + 1
simI[simCt] <- dim(epi)[1]
r <- epi[,2]
i <- epi[,1]
simStore[simCt,1:2,1] <- mle_complete_gsem(r, i, N)
simStore[simCt,3,1] <- simStore[simCt,1,1] / simStore[simCt,2,1]
epi <- sort_gsem(epi)
r <- epi[,2]
simStore[simCt,1:2,2] <- nlm(pbla_partial_gsem, rates, r, N)$estimate
simStore[simCt,3,2] <- simStore[simCt,1,2] / simStore[simCt,2,2]
iepi <- decomplete_gsem(epi, simP, simQ)
iepi <- sort_gsem(iepi)
r <- iepi[,2]
i <- iepi[,1]
simStore[simCt,1:2,3] <- nlm(pbla_incomplete_gsem, rates, r, i, N)$estimate
simStore[simCt,3,3] <- simStore[simCt,1,3] / simStore[simCt,2,3]
if(!(simCt %% simPrint)){print(simCt)}
}
saveRDS(simStore, '../data/vanilla-4-1-200.rds')
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.