Extend qx up to age 105 using HMD regression parameters

$$ logit({}_{5}q_{x+5}) - logit({}_{5}q_x) = \beta_0 + \beta_1 * (logit({}_{5}q_x) - logit({}_{5}q{x-5})) + \text{location}_{RE} $$

Usage

extend_qx_diff(empir_lt, hmd_qx_results, by_vars)

Arguments

empir_lt

data.table with columns: ihme_loc_id, sex, year, age (numeric), mx, ax

hmd_qx_results

data.table with variables sex, age, slope, intercept

by_vars

character vector containing the variable names of all variables that uniquely identify the observations (except for age)

Value

returns empir_lt with the same variables, but modified qx extended to age 105

Details

Recursive prediction