| Library: | eiv |
| See also: | reca |
| Macro: | simex | |
| Description: | SIMEX (SIMulation EXtrapolation) is a simulation-based method of estimating and reducing bias due to measurement error. simex is applicable to general estimation methods, for example, least-squares, maximum likelihood, quasi-likelihood, etc. |
| Usage: | {simexl,simexq} = simex(y, w, z, su2, lam, b) | |
| Input: | ||
| y | n x 1 matrix, the design variables | |
| w | n x 1 matrix, surrogate variable of unobservable variable | |
| z | n x 1 matrix, covariate measured without error | |
| su2 | the variance of the measurement error | |
| lam | pseudo-parameter for generating pseudo-errors | |
| b | the number of replication in each simulation | |
| Output: | ||
| simexl | the estimate based on linear extrapolant function | |
| simexq | the estimate based on quadratic extrapolant function | |
library("xplore")
library("eiv")
n=100
randomize(n)
y=floor(uniform(n)+0.3)
w=uniform(n)^2
z=floor(uniform(n)+0.45)
su2=var(w)/4
lam=aseq(0.01,6,0.5)
b=20
gest=simex(y,w,z,su2,lam,b)
gest.simexl
gest.simexq
Contents of simexl [1,] -2.2569 0.92894 0.90323 Contents of simexq [1,] -2.265 0.95407 0.90822
| Library: | eiv |
| See also: | reca |