Azka Ubaidillah, Ridson Al Farizal P
Ridson Al Farizal P ridsonap@bps.go.id
The goal of sae.projection is to provide tools for performing Small Area Estimation (SAE) using projection-based methods. This approach is particularly effective for generating reliable domain-level estimates in situations where sample sizes are too small for direct estimation. By utilizing auxiliary information and statistical models, the package improves the precision and accuracy of small area estimates.
You can install the development version of sae.projection from GitHub with:
# install.packages("devtools")
devtools::install_github("Alfrzlp/sae.projection")This is a basic example which shows you how to solve a common problem:
library(sae.projection)
library(dplyr)
#> Warning: package 'dplyr' was built under R version 4.2.3
#>
#> Attaching package: 'dplyr'
#> The following objects are masked from 'package:stats':
#>
#> filter, lag
#> The following objects are masked from 'package:base':
#>
#> intersect, setdiff, setequal, uniondf_res <- projection_estimator(
formula = unemployment ~ edu + sex + age + head,
domain = c('prov', 'regency'),
weight = 'weight',
data_model = df_svy2,
data_proj = df_svy1,
model = 'logistic',
cutoff = 'optimum'
)
#> ℹ Getting the optimum cut-off...
#> ✔ The optimum cut-off is 0.105987114406781 and 0.30650406504065
#> Warning in predict_result > cutoff: longer object length is not a multiple of
#> shorter object length
#> # A tibble: 119 × 6
#> prov regency n yhat var_yhat rse
#> <int> <int> <int> <dbl> <dbl> <dbl>
#> 1 31 1 329 0.0633 0.000199 22.3
#> 2 31 71 1229 0.0750 0.0000586 10.2
#> 3 31 72 1368 0.0725 0.0000522 9.97
#> 4 31 73 846 0.0800 0.0000837 11.4
#> 5 31 74 1265 0.109 0.0000714 7.74
#> 6 31 75 1126 0.0941 0.0000633 8.46
#> 7 32 1 1476 0.115 0.0000597 6.71
#> 8 32 2 1224 0.0899 0.0000585 8.50
#> 9 32 3 1248 0.0841 0.0000508 8.48
#> 10 32 4 1353 0.0854 0.0000557 8.73
#> # ℹ 109 more rowsdf_svy2_new <- df_svy2 %>% dplyr::filter(!is.na(income))
m2 <- projection_estimator(
formula = income ~ edu + sex + age + head,
domain = c('prov', 'regency'),
weight = 'weight',
data_model = df_svy2_new,
data_proj = df_svy1,
model = 'linear'
)
#> # A tibble: 119 × 6
#> prov regency n yhat var_yhat rse
#> <int> <int> <int> <dbl> <dbl> <dbl>
#> 1 31 1 329 2579381. 4931305728. 2.72
#> 2 31 71 1229 3202174. 1814357196. 1.33
#> 3 31 72 1368 3158206. 1516683859. 1.23
#> 4 31 73 846 2958848. 2221792017. 1.59
#> 5 31 74 1265 2749878. 1210877276. 1.27
#> 6 31 75 1126 2965990. 1785753160. 1.42
#> 7 32 1 1476 2414944. 876222281. 1.23
#> 8 32 2 1224 2075254. 772421710. 1.34
#> 9 32 3 1248 1996901. 676199395. 1.30
#> 10 32 4 1353 2567675. 1073105001. 1.28
#> # ℹ 109 more rows