mig_estimate_rc {rcbayes}

Estimate Rogers-Castro migration age schedule

Description

Given a set of ages and observed age-specific migrants, estimate the parameters of a Roger-Castro model migration schedule. Choose between a 7, 9, 11 or 13 parameter model.

Usage

mig_estimate_rc(
  ages,
  migrants,
  pop,
  mx,
  sigma,
  pre_working_age,
  working_age,
  retirement,
  post_retirement,
  net_mig,
  ...
)

Arguments

ages

numeric. A vector of integers for ages.

migrants

numeric. A vector of integers for observed age-specific migrants.

pop

numeric. A vector of integers for age-specific population or sample sizes, of which "migrants" experienced a migration event.

mx

numeric. A vector of age-specific migration rates.

sigma

numeric. Standard deviation of migration rates for Normal model. Argument is option, standard deviation is estimated if Normal model is run without being specified.

pre_working_age

logical (TRUE/FALSE). Whether or not to include pre working age component.

working_age

logical (TRUE/FALSE). Whether or not to include working age component.

retirement

logical (TRUE/FALSE). Whether or not to include retirement age component.

post_retirement

logical (TRUE/FALSE). Whether or not to include post retirement age component.

net_mig

numeric. Deprecated argument, use migrants instead.

...

additional inputs to stan, see ?rstan::stan for details.

Value

A list of length 3. The first element, pars_df, is a data frame that provides parameter estimates with 95% credible intervals. The second element, fit_df, is a data frame that shows the data and estimated migration rates at each age. The third element, check_converge, is a data frame that provides the R-hat values and effective sample sizes.

Examples

# Ex 1: Run poisson model using ages, migrants, and population
ages <- 0:80
migrants <- c(202,215,167,188,206,189,164,
            158,197,185,176,173,167,198,
            203,237,249,274,319,345,487,
            491,521,505,529,527,521,529,
            507,484,467,439,399,399,380,
            368,310,324,289,292,270,269,
            285,254,245,265,257,258,263,
            253,346,293,332,346,349,355,
            386,346,344,352,331,320,307,
            320,310,258,254,243,256,263,
            183,169,172,160,166,113,132,
            111,130,110,113)
pop <- c(105505,105505,105505,105505,105505,
        106126,106126,106126,106126,106126,
        100104,100104,100104,100104,100104,
        114880,114880,114880,114880,114880,
        136845,136845,136845,136845,136845,
        136582,136582,136582,136582,136582,
        141935,141935,141935,141935,141935,
        134097,134097,134097,134097,134097,
        130769,130769,130769,130769,130769,
        133718,133718,133718,133718,133718,
        154178,154178,154178,154178,154178,
        145386,145386,145386,145386,145386,
        126270,126270,126270,126270,126270,
        108314,108314,108314,108314,108314,
        79827,79827,79827,79827,79827,59556,
        59556,59556,59556,59556,59556)


# fit the model
res <- mig_estimate_rc(ages = ages, migrants = migrants, pop = pop,
                       pre_working_age = TRUE,
                       working_age = TRUE,
                       retirement = TRUE,
                       post_retirement = FALSE,
                       #optional inputs into stan
                       control = list(adapt_delta = 0.95, max_treedepth = 10),
                       iter = 10, chains = 1 #to speed up example
                       )
# plot the results and data
plot(ages, migrants/pop, ylab = "migration rate", xlab = "age")
lines(ages, res[["fit_df"]]$median, col = "red")
legend("topright", legend=c("data", "fit"), col=c("black", "red"), lty=1, pch = 1)

# Ex 2: Run normal model using ages and mx
ages <- 0:80
mx <- c(0.001914601, 0.002037818, 0.001582863, 0.001781906,
        0.001952514, 0.001780902, 0.001545333, 0.001488796,
        0.001856284, 0.001743211, 0.001758172, 0.001728203,
        0.001668265, 0.001977943, 0.002027891, 0.002063022,
        0.002167479, 0.002385097, 0.002776811, 0.003003134,
        0.003558771, 0.003588001, 0.003807227, 0.003690307,
        0.003865687, 0.003858488, 0.003814558, 0.003873131,
        0.003712056, 0.003543659, 0.003290238, 0.003092965,
        0.002811146, 0.002811146, 0.002677282, 0.002744282,
        0.002311759, 0.002416161, 0.002155156, 0.002177528,
        0.002064710, 0.002057062, 0.002179416, 0.001942356,
        0.001873533, 0.001981783, 0.001921955, 0.001929434,
        0.001966826, 0.001892041, 0.002244159, 0.001900401,
        0.002153355, 0.002244159, 0.002263617, 0.002441776,
        0.002655001, 0.002379872, 0.002366115, 0.002421141,
        0.002621367, 0.002534252, 0.002431298, 0.002534252,
        0.002455057, 0.002381964, 0.002345034, 0.002243477,
        0.002363499, 0.002428126, 0.002292457, 0.002117078,
        0.002154659, 0.002004334, 0.002079497, 0.001897374,
        0.002216401, 0.001863792, 0.002182820, 0.001847001,
        0.001897374)

# fit the model
res <- mig_estimate_rc(ages = ages, mx = mx,
                       pre_working_age = TRUE,
                       working_age = TRUE,
                       retirement = TRUE,
                       post_retirement = FALSE,
                       #optional inputs into stan
                       control = list(adapt_delta = 0.95, max_treedepth = 10),
                       iter = 10, chains = 1 #to speed up example
                       )