Azka Ubaidillah, Ridson Al Farizal P, Silvi Ajeng Larasati, Amelia Rahayu
Ridson Al Farizal P ridsonap@bps.go.id
The sae.projection package provides a robust tool for small area estimation using a projection-based approach. This method is particularly beneficial in scenarios involving two surveys, the first survey collects data solely on auxiliary variables, while the second, typically smaller survey, collects both the variables of interest and the auxiliary variables. The package constructs a working model to predict the variables of interest for each sample in the first survey. These predictions are then used to estimate relevant indicators for the desired domains. This condition overcomes the problem of estimation in a small area when only using the second survey data.
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)
#> Loading required package: tidymodels
#> ── Attaching packages ────────────────────────────────────── tidymodels 1.3.0 ──
#> ✔ broom 1.0.8 ✔ recipes 1.3.1
#> ✔ dials 1.4.0 ✔ rsample 1.3.0
#> ✔ dplyr 1.1.4 ✔ tibble 3.3.0
#> ✔ ggplot2 3.5.2 ✔ tidyr 1.3.1
#> ✔ infer 1.0.8 ✔ tune 1.3.0
#> ✔ modeldata 1.4.0 ✔ workflows 1.2.0
#> ✔ parsnip 1.3.2 ✔ workflowsets 1.1.1
#> ✔ purrr 1.0.4 ✔ yardstick 1.3.2
#> ── Conflicts ───────────────────────────────────────── tidymodels_conflicts() ──
#> ✖ purrr::discard() masks scales::discard()
#> ✖ dplyr::filter() masks stats::filter()
#> ✖ dplyr::lag() masks stats::lag()
#> ✖ recipes::step() masks stats::step()
library(dplyr)df_svy22_income <- df_svy22 %>% filter(!is.na(income))
df_svy23_income <- df_svy23 %>% filter(!is.na(income))lm_proj <- ma_projection(
income ~ age + sex + edu + disability,
cluster_ids = "PSU", weight = "WEIGHT", strata = "STRATA",
domain = c("PROV", "REGENCY"),
working_model = linear_reg(),
data_model = df_svy22_income,
data_proj = df_svy23_income,
nest = TRUE
)
lm_proj$projection
#> PROV REGENCY ypr var_ypr rse_ypr
#> 1 40 1 2355859 112850851251 14.259463
#> 2 40 2 2155687 14070372903 5.502588
#> 3 40 3 2088804 9711460386 4.717855
#> 4 40 4 2188828 12511600884 5.110281
#> 5 40 5 2138723 22768231513 7.055213
#> 6 40 6 2061848 10210621315 4.900827
#> 7 40 7 2085265 9976295858 4.789866
#> 8 40 8 2028248 6536668678 3.986183
#> 9 40 9 2233891 12561471412 5.017163
#> 10 40 10 2997192 470694250001 22.890476
#> 11 40 11 2061056 10060073816 4.866433
#> 12 40 12 2201031 9581460318 4.447231
#> 13 40 13 2213420 32332172691 8.123695
#> 14 40 14 2200890 13896011417 5.356075
#> 15 40 15 2269106 15294166332 5.450146
#> 16 40 16 2174782 12143176114 5.066994
#> 17 40 17 1965337 10785195069 5.284172
#> 18 40 18 1837433 8405642799 4.989694
#> 19 40 19 2208518 24478147649 7.084160
#> 20 40 20 2143794 13056774386 5.330094
#> 21 40 21 2067066 7984452688 4.322831
#> 22 40 22 2038540 9802389037 4.856761
#> 23 40 23 2393420 86969130589 12.321503
#> 24 40 24 2240375 20747721026 6.429313
#> 25 40 25 2195086 14570971410 5.499113
#> 26 40 26 2187382 24209446221 7.113247
#> 27 40 27 2217716 64407458495 11.443585
#> 28 40 28 1986641 6921930259 4.187880
#> 29 40 29 1965301 5897858683 3.907672
#> 30 40 30 2256849 17114638305 5.796707
#> 31 40 31 2267756 24897436539 6.957945
#> 32 40 32 2063165 8403451069 4.443189
#> 33 40 33 1982999 7567172983 4.386765
#> 34 40 34 2244054 21199661590 6.488302
#> 35 40 35 2137450 38140066571 9.136813
#> 36 40 36 2024760 10698498794 5.108435
#> 37 40 37 2055319 18930473132 6.694245
#> 38 40 38 2954545 314127308067 18.969774
#> 39 40 39 1955815 13455721680 5.930971
#> 40 40 40 1998518 23669511858 7.698154
#> 41 40 41 2387539 63038731173 10.516066
#> 42 40 42 2007678 8297754980 4.537183
#> 43 40 43 1983971 5647857868 3.787971
#> 44 40 44 2187359 41458503574 9.308653
#> 45 40 45 1986850 14801246261 6.123279
#> 46 40 46 2369022 48861397759 9.330694
#> 47 40 47 2116854 9339080301 4.565214
#> 48 40 48 2174280 19690945724 6.453833
#> 49 40 49 2239364 41590806935 9.106972
#> 50 40 50 2118603 13019855312 5.385840
#> 51 40 51 1927953 6732676376 4.255959
#> 52 40 52 1970130 7315602106 4.341404
#> 53 40 53 1954391 7289520282 4.368556
#> 54 40 54 2137444 15690069691 5.860272
#> 55 40 55 2017267 6428930853 3.974715
#> 56 40 56 2351050 50853300343 9.591748
#> 57 40 57 2162501 25747523120 7.420129
#> 58 40 58 2674530 313400377469 20.931591
#> 59 40 59 2335474 17471370430 5.659635
#> 60 40 60 2120891 9692679744 4.641983
#> 61 40 61 2098821 13706060975 5.578031
#> 62 40 62 1971450 9299512082 4.891525
#> 63 40 63 2209349 32676097067 8.181835
#> 64 40 64 1941592 6738894142 4.228014
#> 65 40 65 2070572 7158546619 4.086223
#> 66 40 66 2089372 17862942632 6.396771
#> 67 40 67 1900122 7741749874 4.630608
#> 68 40 68 1941208 6187014231 4.051992
#> 69 40 69 2279011 33956910517 8.085701
#> 70 40 70 2142969 14310388261 5.582257
#> 71 40 71 2341867 30248638021 7.426611
#> 72 40 72 1968031 7531981310 4.409839
#> 73 40 73 2333145 31242426098 7.575837
#> 74 40 74 2191866 15913071184 5.755235
#> 75 40 75 2390506 52976205496 9.628322
#> 76 40 76 1937820 6596212084 4.191157
#> 77 40 77 2057303 20972153977 7.039198
#> 78 40 78 2269991 34769289337 8.214362
#> 79 40 79 2166598 17221385107 6.056974df_svy22_neet <- df_svy22 %>%
filter(between(age, 15, 24))
df_svy23_neet <- df_svy23 %>%
filter(between(age, 15, 24))lr_proj <- ma_projection(
formula = neet ~ sex + edu + disability,
cluster_ids = "PSU",
weight = "WEIGHT",
strata = "STRATA",
domain = c("PROV", "REGENCY"),
working_model = logistic_reg(),
data_model = df_svy22_neet,
data_proj = df_svy23_neet,
nest = TRUE
)
lr_proj$projection
#> PROV REGENCY ypr var_ypr rse_ypr
#> 1 40 1 1.1183886 0.006444221 7.177819
#> 2 40 2 1.2557005 0.006105655 6.222717
#> 3 40 3 1.1274297 0.006065054 6.907611
#> 4 40 4 1.1113317 0.005828028 6.869373
#> 5 40 5 1.0800544 0.006010325 7.177998
#> 6 40 6 1.1517670 0.006139404 6.802969
#> 7 40 7 1.0794256 0.007007080 7.754894
#> 8 40 8 1.2443838 0.005809545 6.125150
#> 9 40 9 1.2047362 0.004965885 5.849333
#> 10 40 10 0.9640855 0.006984372 8.668583
#> 11 40 11 1.0933346 0.005826673 6.981636
#> 12 40 12 1.1801704 0.005316427 6.178246
#> 13 40 13 1.1846172 0.007200787 7.163280
#> 14 40 14 1.0845802 0.006903031 7.660520
#> 15 40 15 1.1505433 0.006664377 7.095399
#> 16 40 16 1.1486635 0.005427596 6.413734
#> 17 40 17 1.2093561 0.005661401 6.221681
#> 18 40 18 1.1028660 0.006612135 7.373066
#> 19 40 19 1.2294521 0.004804973 5.638114
#> 20 40 20 1.1578711 0.006367011 6.891402
#> 21 40 21 1.2762862 0.005070637 5.579345
#> 22 40 22 1.2102351 0.005540878 6.150629
#> 23 40 23 1.2059524 0.005400307 6.093672
#> 24 40 24 1.1783683 0.005738102 6.428402
#> 25 40 25 1.1080694 0.005155669 6.480008
#> 26 40 26 1.2509160 0.005157540 5.741074
#> 27 40 27 1.2471413 0.005605561 6.003353
#> 28 40 28 1.1753644 0.004616937 5.781020
#> 29 40 29 1.0877470 0.005710199 6.947007
#> 30 40 30 1.0443887 0.006598061 7.777607
#> 31 40 31 1.2117605 0.006043338 6.415369
#> 32 40 32 1.1688809 0.005417636 6.297014
#> 33 40 33 1.1884573 0.006069021 6.555046
#> 34 40 34 1.1248072 0.006022908 6.899618
#> 35 40 35 1.1672254 0.006522315 6.919045
#> 36 40 36 1.2252635 0.005834061 6.233845
#> 37 40 37 1.0434104 0.006060357 7.460947
#> 38 40 38 1.2600382 0.004835964 5.518967
#> 39 40 39 1.1189457 0.006639541 7.282157
#> 40 40 40 1.0950714 0.005302188 6.649441
#> 41 40 41 1.1898514 0.005923444 6.468363
#> 42 40 42 1.2487870 0.007639726 6.999234
#> 43 40 43 1.0982993 0.006348754 7.254768
#> 44 40 44 1.1818942 0.005524671 6.288899
#> 45 40 45 1.2766450 0.006564731 6.346559
#> 46 40 46 1.1639300 0.005717720 6.496577
#> 47 40 47 1.2213804 0.006462049 6.581641
#> 48 40 48 1.1945465 0.005625430 6.278773
#> 49 40 49 1.0425183 0.005626184 7.194876
#> 50 40 50 0.9518069 0.007261626 8.952989
#> 51 40 51 1.3295578 0.005528923 5.592591
#> 52 40 52 1.0509247 0.005438234 7.017095
#> 53 40 53 1.1885369 0.005763973 6.387756
#> 54 40 54 1.1534033 0.005731217 6.563602
#> 55 40 55 1.2341504 0.005928711 6.238958
#> 56 40 56 1.1968135 0.006673016 6.825502
#> 57 40 57 1.1810190 0.005667091 6.374164
#> 58 40 58 1.1206938 0.005539195 6.641044
#> 59 40 59 1.0108983 0.005954844 7.633571
#> 60 40 60 1.1346988 0.006363699 7.030306
#> 61 40 61 1.2145020 0.005925792 6.338332
#> 62 40 62 1.1014442 0.005650843 6.824866
#> 63 40 63 1.2185587 0.005020100 5.814464
#> 64 40 64 1.1159900 0.005480422 6.633561
#> 65 40 65 1.1039627 0.005699648 6.838638
#> 66 40 66 1.1283841 0.006518595 7.155172
#> 67 40 67 1.0386828 0.005776468 7.317256
#> 68 40 68 0.9854114 0.005870607 7.775421
#> 69 40 69 1.1610866 0.005587405 6.437844
#> 70 40 70 1.1255552 0.006023267 6.895238
#> 71 40 71 1.2091412 0.005741644 6.266732
#> 72 40 72 1.2197376 0.005334661 5.988072
#> 73 40 73 1.1532298 0.005458433 6.406460
#> 74 40 74 1.1886667 0.006554084 6.810765
#> 75 40 75 1.2285712 0.005805988 6.202085
#> 76 40 76 1.1290612 0.006445259 7.110542
#> 77 40 77 1.1334380 0.005447368 6.511718
#> 78 40 78 1.0864061 0.006680768 7.523519
#> 79 40 79 1.1952840 0.004899836 5.856251library(bonsai)
show_engines("boost_tree")
#> # A tibble: 7 × 2
#> engine mode
#> <chr> <chr>
#> 1 xgboost classification
#> 2 xgboost regression
#> 3 C5.0 classification
#> 4 spark classification
#> 5 spark regression
#> 6 lightgbm regression
#> 7 lightgbm classification
lgbm_model <- boost_tree(
mtry = tune(), trees = tune(), min_n = tune(), tree_depth = tune(), learn_rate = tune(),
engine = "lightgbm"
)lgbm_proj <- ma_projection(
formula = neet ~ sex + edu + disability,
cluster_ids = "PSU",
weight = "WEIGHT",
strata = "STRATA",
domain = c("PROV", "REGENCY"),
working_model = lgbm_model,
data_model = df_svy22_neet,
data_proj = df_svy23_neet,
cv_folds = 3,
grid = 20,
nest = TRUE
)
lgbm_proj$projectiondata(df_svy_A)
data(df_svy_B)
x_predictors <- names(df_svy_A)[5:19]
# Projection RF with feature selection and bias correction
result <- projection_randomforest(
data_model = df_svy_A,
target_column = "Y",
predictor_cols = x_predictors,
data_proj = df_svy_B,
domain1 = "province",
domain2 = "regency",
psu = "num",
ssu = NULL,
strata = NULL,
weights = "weight",
feature_selection = TRUE,
bias_correction = TRUE
)
#> Info: Bias correction is enabled. Calculating indirect estimation with bias
#> correction.
#> Starting preprocessing...
#> Preprocessing completed. Starting data split...
#> Data split completed. Checking for missing values...
#> Feature selection (RFE) enabled. Starting RFE...
#> Loading required package: lattice
#>
#>
#> Attaching package: 'caret'
#>
#>
#> The following objects are masked from 'package:yardstick':
#>
#> precision, recall, sensitivity, specificity
#>
#>
#> The following object is masked from 'package:purrr':
#>
#> lift
#>
#>
#> RFE completed. Features selected.
#> Features selected. Starting hyperparameter tuning...
#> Hyperparameter tuning completed. Training model...
#> note: only 7 unique complexity parameters in default grid. Truncating the grid to 7 .
#> Model training completed. Starting model evaluation...
#> Evaluation completed. Starting prediction on new data...
#> Prediction completed. Starting indirect estimation for domain...
#> Indirect estimation completed. Starting direct estimation...
#> Direct estimation completed. Starting bias correction estimation...
#> Bias-corrected estimation completed. Returning results...
print(result)
#> $model
#> Random Forest
#>
#> 1600 samples
#> 8 predictor
#> 2 classes: 'No', 'Yes'
#>
#> No pre-processing
#> Resampling: Cross-Validated (5 fold)
#> Summary of sample sizes: 1280, 1281, 1279, 1280, 1280
#> Addtional sampling using SMOTE
#>
#> Resampling results across tuning parameters:
#>
#> mtry splitrule Accuracy Kappa
#> 2 gini 0.8918628 0.7596156
#> 2 extratrees 0.8818569 0.7374146
#> 3 gini 0.8912457 0.7583682
#> 3 extratrees 0.8843667 0.7431274
#> 4 gini 0.8868706 0.7489424
#> 4 extratrees 0.8831167 0.7407826
#> 5 gini 0.8837456 0.7425685
#> 5 extratrees 0.8824878 0.7394682
#> 6 gini 0.8831226 0.7413135
#> 6 extratrees 0.8831167 0.7411973
#> 7 gini 0.8849937 0.7455293
#> 7 extratrees 0.8843667 0.7439495
#> 8 gini 0.8800015 0.7347802
#> 8 extratrees 0.8818648 0.7388603
#>
#> Tuning parameter 'min.node.size' was held constant at a value of 1
#> Accuracy was used to select the optimal model using the largest value.
#> The final values used for the model were mtry = 2, splitrule = gini
#> and min.node.size = 1.
#>
#> $importance
#> ranger variable importance
#>
#> Overall
#> x10 100.000
#> x4 33.334
#> x6 25.510
#> x1 19.271
#> x2 19.219
#> x8 18.952
#> x5 3.924
#> x12 0.000
#>
#> $train_accuracy
#> Accuracy
#> 0.95125
#>
#> $validation_accuracy
#> Accuracy
#> 0.915
#>
#> $validation_performance
#> Confusion Matrix and Statistics
#>
#> Reference
#> Prediction No Yes
#> No 121 11
#> Yes 23 245
#>
#> Accuracy : 0.915
#> 95% CI : (0.8832, 0.9404)
#> No Information Rate : 0.64
#> P-Value [Acc > NIR] : < 2e-16
#>
#> Kappa : 0.8121
#>
#> Mcnemar's Test P-Value : 0.05923
#>
#> Sensitivity : 0.8403
#> Specificity : 0.9570
#> Pos Pred Value : 0.9167
#> Neg Pred Value : 0.9142
#> Prevalence : 0.3600
#> Detection Rate : 0.3025
#> Detection Prevalence : 0.3300
#> Balanced Accuracy : 0.8987
#>
#> 'Positive' Class : No
#>
#>
#> $data_proj
#> # A tibble: 8,000 × 21
#> province regency id_ind num weight x1 x2 x3 x4 x5 x6
#> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl>
#> 1 31 19 18362 1 5 0 2 1 70 3 9
#> 2 31 25 24661 2 5 1 2 1 35 4 1
#> 3 31 30 29630 3 5 0 2 0 57 5 6
#> 4 31 18 17536 4 5 1 1 0 49 2 3
#> 5 31 2 1967 5 5 1 2 0 45 5 4
#> 6 31 14 13395 6 5 1 1 1 42 1 5
#> 7 31 5 4673 7 5 0 2 1 19 4 9
#> 8 31 27 26950 8 5 0 2 0 39 4 10
#> 9 31 5 4701 9 5 1 1 1 40 3 1
#> 10 31 22 21696 10 5 0 1 1 64 2 4
#> # ℹ 7,990 more rows
#> # ℹ 10 more variables: x7 <dbl>, x8 <dbl>, x9 <dbl>, x10 <dbl>, x11 <dbl>,
#> # x12 <dbl>, x13 <dbl>, x14 <dbl>, x15 <dbl>, Est_Y <fct>
#>
#> $Direct
#> Estimation_Direct Variance RSE
#> 31 64 0.000115 1.68
#>
#> $Domain1_corrected_bias
#> Estimation_Domain1 Variance RSE Est_corrected Var_corrected RSE_corrected
#> 31 68.74 2.7e-05 0.75 65.84 3.2e-05 0.86
#>
#> $Domain2_corrected_bias
#> Estimation_Domain2 Variance RSE Est_corrected Var_corrected
#> 31.1 66.30 0.001214 5.26 63.40 0.001440
#> 31.2 71.69 0.000927 4.25 68.79 0.001153
#> 31.3 68.61 0.000966 4.53 65.71 0.001192
#> 31.4 64.29 0.001093 5.14 61.39 0.001319
#> 31.5 64.21 0.001210 5.42 61.31 0.001436
#> 31.6 74.44 0.001057 4.37 71.54 0.001283
#> 31.7 66.03 0.001073 4.96 63.13 0.001299
#> 31.8 68.35 0.000992 4.61 65.45 0.001218
#> 31.9 57.87 0.001238 6.08 54.97 0.001464
#> 31.10 79.17 0.000764 3.49 76.27 0.000990
#> 31.11 70.17 0.001157 4.85 67.27 0.001383
#> 31.12 67.84 0.000961 4.57 64.94 0.001187
#> 31.13 67.19 0.001148 5.04 64.29 0.001374
#> 31.14 66.18 0.001081 4.97 63.28 0.001307
#> 31.15 65.70 0.001089 5.02 62.80 0.001315
#> 31.16 72.19 0.001188 4.77 69.29 0.001414
#> 31.17 65.32 0.001310 5.54 62.42 0.001536
#> 31.18 69.07 0.001101 4.80 66.17 0.001327
#> 31.19 71.65 0.001047 4.52 68.75 0.001273
#> 31.20 66.19 0.001066 4.93 63.29 0.001292
#> 31.21 77.08 0.000920 3.94 74.18 0.001146
#> 31.22 66.67 0.001043 4.85 63.77 0.001269
#> 31.23 73.04 0.000965 4.25 70.14 0.001191
#> 31.24 65.48 0.001148 5.17 62.58 0.001374
#> 31.25 74.09 0.000995 4.26 71.19 0.001221
#> 31.26 70.14 0.000993 4.49 67.24 0.001219
#> 31.27 68.87 0.001011 4.62 65.97 0.001237
#> 31.28 66.50 0.001098 4.98 63.60 0.001324
#> 31.29 69.61 0.001037 4.63 66.71 0.001263
#> 31.30 68.04 0.001121 4.92 65.14 0.001347
#> 31.31 63.29 0.001123 5.29 60.39 0.001349
#> 31.32 69.54 0.001075 4.72 66.64 0.001301
#> 31.33 69.89 0.001131 4.81 66.99 0.001357
#> 31.34 70.24 0.001020 4.55 67.34 0.001246
#> 31.35 68.11 0.001174 5.03 65.21 0.001400
#> 31.36 68.29 0.001056 4.76 65.39 0.001282
#> 31.37 69.00 0.001070 4.74 66.10 0.001296
#> 31.38 71.57 0.001033 4.49 68.67 0.001259
#> 31.39 70.33 0.000998 4.49 67.43 0.001224
#> 31.40 67.74 0.001175 5.06 64.84 0.001401
#> RSE_corrected
#> 31.1 5.99
#> 31.2 4.94
#> 31.3 5.25
#> 31.4 5.92
#> 31.5 6.18
#> 31.6 5.01
#> 31.7 5.71
#> 31.8 5.33
#> 31.9 6.96
#> 31.10 4.13
#> 31.11 5.53
#> 31.12 5.31
#> 31.13 5.77
#> 31.14 5.71
#> 31.15 5.77
#> 31.16 5.43
#> 31.17 6.28
#> 31.18 5.51
#> 31.19 5.19
#> 31.20 5.68
#> 31.21 4.56
#> 31.22 5.59
#> 31.23 4.92
#> 31.24 5.92
#> 31.25 4.91
#> 31.26 5.19
#> 31.27 5.33
#> 31.28 5.72
#> 31.29 5.33
#> 31.30 5.63
#> 31.31 6.08
#> 31.32 5.41
#> 31.33 5.50
#> 31.34 5.24
#> 31.35 5.74
#> 31.36 5.48
#> 31.37 5.45
#> 31.38 5.17
#> 31.39 5.19
#> 31.40 5.77data(df_svy_A)
data(df_svy_B)
# Remove unused identifier column
df_svy_A <- df_svy_A |> dplyr::select(-id_ind)
df_svy_B <- df_svy_B |> dplyr::select(-id_ind)
# Projection XGBoost with feature selection and bias correction
result <- projection_xgboost(
target_col = "Y",
data_model = df_svy_A,
data_proj = df_svy_B,
id = "num",
STRATA = NULL,
domain1 = "province",
domain2 = "regency",
weight = "weight",
nfold = 10,
test_size = 0.2,
task_type = "classification",
feature_selection = TRUE,
corrected_bias = TRUE
)
#>
#> Load data...
#>
#> Final Task Type: binary
#>
#> Handling missing values in preprocessing...
#>
#> Performing feature selection...
#>
#> Applying SMOTE to balance classes...
#>
#> Preparing data for XGBoost...
#>
#> Performing hyperparameter tuning...
#>
#> Evaluating model...
#>
#> Performing survey estimation...
#>
#> Computing Corrected Bias...
#>
#> Computing Direct Estimate...
#>
#> PSU, SSU, and STRATA are available. Proceeding with the data...
#>
#> Corrected Bias...
#>
#> Weight Index...
#>
#> Bias...
#>
#> Corrected Bias Calculation Completed...
print(result)
#> $metadata
#> $metadata$method
#> [1] "Projection Estimator With XGBoost Algorithm"
#>
#> $metadata$model_type
#> [1] "classification"
#>
#> $metadata$feature_selection_used
#> [1] TRUE
#>
#> $metadata$corrected_bias_applied
#> [1] TRUE
#>
#> $metadata$n_features_used
#> [1] 6
#>
#> $metadata$model_params
#> $metadata$model_params$objective
#> [1] "binary:logistic"
#>
#> $metadata$model_params$eval_metric
#> [1] "logloss"
#>
#> $metadata$model_params$eta
#> [1] 0.1
#>
#> $metadata$model_params$max_depth
#> [1] 7
#>
#> $metadata$model_params$min_child_weight
#> [1] 1
#>
#> $metadata$model_params$subsample
#> [1] 0.8
#>
#> $metadata$model_params$colsample_bytree
#> [1] 0.7
#>
#> $metadata$model_params$lambda
#> [1] 0
#>
#> $metadata$model_params$alpha
#> [1] 0
#>
#> $metadata$model_params$nthread
#> [1] 4
#>
#> $metadata$model_params$validate_parameters
#> [1] TRUE
#>
#>
#> $metadata$features_selected
#> [1] "x1" "x2" "x4" "x6" "x10" "x11"
#>
#>
#> $estimation
#> $estimation$projected_data
#> # A tibble: 8,000 × 20
#> province regency num weight x1 x2 x3 x4 x5 x6 x7 x8
#> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl>
#> 1 31 19 1 5 0 2 1 70 3 9 0 429
#> 2 31 25 2 5 1 2 1 35 4 1 0 469
#> 3 31 30 3 5 0 2 0 57 5 6 0 95
#> 4 31 18 4 5 1 1 0 49 2 3 0 295
#> 5 31 2 5 5 1 2 0 45 5 4 0 129
#> 6 31 14 6 5 1 1 1 42 1 5 0 269
#> 7 31 5 7 5 0 2 1 19 4 9 1 172
#> 8 31 27 8 5 0 2 0 39 4 10 1 555
#> 9 31 5 9 5 1 1 1 40 3 1 0 150
#> 10 31 22 10 5 0 1 1 64 2 4 0 277
#> # ℹ 7,990 more rows
#> # ℹ 8 more variables: x9 <dbl>, x10 <dbl>, x11 <dbl>, x12 <dbl>, x13 <dbl>,
#> # x14 <dbl>, x15 <dbl>, P.hat <dbl>
#>
#> $estimation$domain1_estimation
#> province Estimation RSE Variance
#> 31 31 67.77 0.77 2.730403e-05
#>
#> $estimation$domain2_estimation
#> province regency Estimation RSE Variance
#> 31.1 31 1 65.22 5.38 0.0012329971
#> 31.2 31 2 71.69 4.25 0.0009268582
#> 31.3 31 3 67.26 4.67 0.0009875380
#> 31.4 31 4 63.81 5.20 0.0010998027
#> 31.5 31 5 63.68 5.48 0.0012173850
#> 31.6 31 6 73.33 4.49 0.0010865556
#> 31.7 31 7 64.59 5.12 0.0010944122
#> 31.8 31 8 68.81 4.56 0.0009846570
#> 31.9 31 9 57.36 6.14 0.0012416905
#> 31.10 31 10 77.78 3.64 0.0008002829
#> 31.11 31 11 71.82 4.66 0.0011182325
#> 31.12 31 12 67.40 4.62 0.0009680546
#> 31.13 31 13 67.19 5.04 0.0011483674
#> 31.14 31 14 63.77 5.24 0.0011162936
#> 31.15 31 15 63.77 5.24 0.0011162936
#> 31.16 31 16 71.01 4.92 0.0012183484
#> 31.17 31 17 64.74 5.61 0.0013196654
#> 31.18 31 18 67.53 4.98 0.0011304750
#> 31.19 31 19 69.59 4.75 0.0010910254
#> 31.20 31 20 66.19 4.93 0.0010657849
#> 31.21 31 21 74.48 4.22 0.0009901083
#> 31.22 31 22 65.73 4.95 0.0010577094
#> 31.23 31 23 72.55 4.31 0.0009763680
#> 31.24 31 24 64.47 5.29 0.0011629402
#> 31.25 31 25 72.02 4.49 0.0010442174
#> 31.26 31 26 69.19 4.59 0.0010103530
#> 31.27 31 27 69.81 4.52 0.0009942335
#> 31.28 31 28 67.00 4.93 0.0010893812
#> 31.29 31 29 69.12 4.68 0.0010464620
#> 31.30 31 30 65.98 5.16 0.0011571856
#> 31.31 31 31 60.87 5.57 0.0011507973
#> 31.32 31 32 69.04 4.77 0.0010852364
#> 31.33 31 33 69.89 4.81 0.0011314799
#> 31.34 31 34 67.80 4.81 0.0010650045
#> 31.35 31 35 64.32 5.48 0.0012405949
#> 31.36 31 36 68.29 4.76 0.0010564139
#> 31.37 31 37 68.00 4.85 0.0010881360
#> 31.38 31 38 70.05 4.66 0.0010650910
#> 31.39 31 39 69.38 4.60 0.0010166311
#> 31.40 31 40 66.67 5.19 0.0011948925
#>
#>
#> $performance
#> $performance$mean_train_accuracy
#> [1] 0.9324417
#>
#> $performance$final_accuracy
#> Accuracy
#> 0.8919271
#>
#> $performance$confusion_matrix
#> Confusion Matrix and Statistics
#>
#> Reference
#> Prediction 0 1
#> 0 313 41
#> 1 42 372
#>
#> Accuracy : 0.8919
#> 95% CI : (0.8678, 0.913)
#> No Information Rate : 0.5378
#> P-Value [Acc > NIR] : <2e-16
#>
#> Kappa : 0.7826
#>
#> Mcnemar's Test P-Value : 1
#>
#> Sensitivity : 0.8817
#> Specificity : 0.9007
#> Pos Pred Value : 0.8842
#> Neg Pred Value : 0.8986
#> Prevalence : 0.4622
#> Detection Rate : 0.4076
#> Detection Prevalence : 0.4609
#> Balanced Accuracy : 0.8912
#>
#> 'Positive' Class : 0
#>
#>
#>
#> $bias_correction
#> $bias_correction$direct_estimation
#> province Estimation RSE Variance
#> 31 31 64 1.68 0.0001152576
#>
#> $bias_correction$corrected_domain1
#> # A tibble: 1 × 4
#> province Est_corrected Var_corrected RSE_corrected
#> <dbl> <dbl> <dbl> <dbl>
#> 1 31 65.1 0.000033 0.88
#>
#> $bias_correction$corrected_domain2
#> province regency Estimation RSE Variance bias_est var_bias_est
#> 1 31 1 65.22 5.38 0.0012329971 -2.65 0.0002264521
#> 2 31 2 71.69 4.25 0.0009268582 -2.65 0.0002264521
#> 3 31 3 67.26 4.67 0.0009875380 -2.65 0.0002264521
#> 4 31 4 63.81 5.20 0.0010998027 -2.65 0.0002264521
#> 5 31 5 63.68 5.48 0.0012173850 -2.65 0.0002264521
#> 6 31 6 73.33 4.49 0.0010865556 -2.65 0.0002264521
#> 7 31 7 64.59 5.12 0.0010944122 -2.65 0.0002264521
#> 8 31 8 68.81 4.56 0.0009846570 -2.65 0.0002264521
#> 9 31 9 57.36 6.14 0.0012416905 -2.65 0.0002264521
#> 10 31 10 77.78 3.64 0.0008002829 -2.65 0.0002264521
#> 11 31 11 71.82 4.66 0.0011182325 -2.65 0.0002264521
#> 12 31 12 67.40 4.62 0.0009680546 -2.65 0.0002264521
#> 13 31 13 67.19 5.04 0.0011483674 -2.65 0.0002264521
#> 14 31 14 63.77 5.24 0.0011162936 -2.65 0.0002264521
#> 15 31 15 63.77 5.24 0.0011162936 -2.65 0.0002264521
#> 16 31 16 71.01 4.92 0.0012183484 -2.65 0.0002264521
#> 17 31 17 64.74 5.61 0.0013196654 -2.65 0.0002264521
#> 18 31 18 67.53 4.98 0.0011304750 -2.65 0.0002264521
#> 19 31 19 69.59 4.75 0.0010910254 -2.65 0.0002264521
#> 20 31 20 66.19 4.93 0.0010657849 -2.65 0.0002264521
#> 21 31 21 74.48 4.22 0.0009901083 -2.65 0.0002264521
#> 22 31 22 65.73 4.95 0.0010577094 -2.65 0.0002264521
#> 23 31 23 72.55 4.31 0.0009763680 -2.65 0.0002264521
#> 24 31 24 64.47 5.29 0.0011629402 -2.65 0.0002264521
#> 25 31 25 72.02 4.49 0.0010442174 -2.65 0.0002264521
#> 26 31 26 69.19 4.59 0.0010103530 -2.65 0.0002264521
#> 27 31 27 69.81 4.52 0.0009942335 -2.65 0.0002264521
#> 28 31 28 67.00 4.93 0.0010893812 -2.65 0.0002264521
#> 29 31 29 69.12 4.68 0.0010464620 -2.65 0.0002264521
#> 30 31 30 65.98 5.16 0.0011571856 -2.65 0.0002264521
#> 31 31 31 60.87 5.57 0.0011507973 -2.65 0.0002264521
#> 32 31 32 69.04 4.77 0.0010852364 -2.65 0.0002264521
#> 33 31 33 69.89 4.81 0.0011314799 -2.65 0.0002264521
#> 34 31 34 67.80 4.81 0.0010650045 -2.65 0.0002264521
#> 35 31 35 64.32 5.48 0.0012405949 -2.65 0.0002264521
#> 36 31 36 68.29 4.76 0.0010564139 -2.65 0.0002264521
#> 37 31 37 68.00 4.85 0.0010881360 -2.65 0.0002264521
#> 38 31 38 70.05 4.66 0.0010650910 -2.65 0.0002264521
#> 39 31 39 69.38 4.60 0.0010166311 -2.65 0.0002264521
#> 40 31 40 66.67 5.19 0.0011948925 -2.65 0.0002264521
#> Est_corrected Var_corrected RSE_corrected
#> 1 62.57 0.001459 6.10
#> 2 69.04 0.001153 4.92
#> 3 64.61 0.001214 5.39
#> 4 61.16 0.001326 5.95
#> 5 61.03 0.001444 6.23
#> 6 70.68 0.001313 5.13
#> 7 61.94 0.001321 5.87
#> 8 66.16 0.001211 5.26
#> 9 54.71 0.001468 7.00
#> 10 75.13 0.001027 4.27
#> 11 69.17 0.001345 5.30
#> 12 64.75 0.001195 5.34
#> 13 64.54 0.001375 5.75
#> 14 61.12 0.001343 6.00
#> 15 61.12 0.001343 6.00
#> 16 68.36 0.001445 5.56
#> 17 62.09 0.001546 6.33
#> 18 64.88 0.001357 5.68
#> 19 66.94 0.001317 5.42
#> 20 63.54 0.001292 5.66
#> 21 71.83 0.001217 4.86
#> 22 63.08 0.001284 5.68
#> 23 69.90 0.001203 4.96
#> 24 61.82 0.001389 6.03
#> 25 69.37 0.001271 5.14
#> 26 66.54 0.001237 5.29
#> 27 67.16 0.001221 5.20
#> 28 64.35 0.001316 5.64
#> 29 66.47 0.001273 5.37
#> 30 63.33 0.001384 5.87
#> 31 58.22 0.001377 6.37
#> 32 66.39 0.001312 5.46
#> 33 67.24 0.001358 5.48
#> 34 65.15 0.001291 5.52
#> 35 61.67 0.001467 6.21
#> 36 65.64 0.001283 5.46
#> 37 65.35 0.001315 5.55
#> 38 67.40 0.001292 5.33
#> 39 66.73 0.001243 5.28
#> 40 64.02 0.001421 5.89