merge_states: Marginalizing over states.

In niklasmaltzahn/multistate: simulation and estimation of multistate models

Description

Merging states in a multistate model to produce new (marginal) model.

Usage

merge_states(df, new_states, new_tmat)

Arguments

df	Multistate data frame of the type from sim_fun.
new_states	A list list of how new states are composed of old states.
new_tmat	The corresponding tmat (see tmat_function) for the marginal multistate model.

Examples

jump_types <- matrix(c(NA, "exp", "exp", NA, "exp",
                       "exp", NA ,NA,"exp", "exp",
                       "exp", NA, NA, "exp", "exp",
                       NA, "exp", "exp", NA, "exp",
                       rep(NA, times = 5)), 5, 5, byrow = T)
tmat <- tmat_function(jump_types)

lambda <- matrix(c(NA, 0.1, 0.15, NA, 0.08,
                   0.12, NA ,NA,0.15, 0.05,
                   0.1, NA, NA, 0.1, 0.08,
                   NA, 0.12, 0.1, NA, 0.05,
                   rep(NA, times = 5)), 5, 5, byrow = T)

par_list2 <- list(lambda = lambda)

data_sim <- sim_fun(init_probs = c(0.5,0,0.5,rep(0, times = 2)), jump_types, pars = par_list2,
                   cens_time = 1000, nbr = 1000)


new_states <- list(state1 = 1:4, state2 = 5)
    new_tmat <- matrix(NA, 2, 2)
    new_tmat[1,2] <- 1

data_margial <- merge_states(data_sim, new_states, new_tmat)

niklasmaltzahn/multistate documentation built on Nov. 4, 2019, 10:06 p.m.