Hybrid Fay-Herriot Random Forest for Small Area Estimation of Per Capita Expenditure in Bandung Regency

Nabillah Rahmatiah Tangke; Asep Saefuddin; Anang Kurnia

doi:10.31764/jtam.v10i4.39931

Hybrid Fay-Herriot Random Forest for Small Area Estimation of Per Capita Expenditure in Bandung Regency

Nabillah Rahmatiah Tangke, Asep Saefuddin, Anang Kurnia

Abstract

Direct survey estimates for small areas tend to be too noisy to use on their own when only 15 to 20 households are sampled per area, which is the situation at the sub-district level in Bandung Regency. The Fay-Herriot model in Small Area Estimation (SAE) helps by borrowing strength across areas, but it leans on two assumptions that do not hold well for sub-district per capita expenditure: linearity in the auxiliary variables, and normality of the sampling errors and the area random effects. In practice, PODES indicators relate to expenditure through thresholds and interactions rather than in a straight line, and per capita expenditure at the sub-district level is right-skewed. Both are problems for the classic Fay-Herriot model, and we address them in a single model. Methodologically, it is a quantitative, model-based comparative study that aims to develop and evaluate a hybrid area-level estimator that relaxes the linearity assumption of the Fay-Herriot model and to assess its accuracy for estimating sub-district per capita expenditure. The linear fixed effect is replaced by a Random Forest, yielding a hybrid Fay-Herriot Random Forest (FH-RF), and the response is log-transformed with a Slud and Maiti style bias correction when predictions are mapped back to the original scale. The model is applied to average weekly per capita expenditure on food, beverages, and tobacco across all 31 sub-districts of Bandung Regency, with the response drawn from the 2023 Human Development Index Survey and ten auxiliary variables from PODES. Eight scenarios were compared, varying the model type, the response transformation, and the source of sampling variance. The hybrid FH-RF Log-Normal model with GVF-smoothed variance performed best, with a mean RRMSE of 7.14%, against 9.05% for the strongest classic Fay-Herriot model and 12.75% for the direct estimate, a 44.0% improvement over direct estimation and 21.1% over the best classic specification. 28 of the 31 sub-districts (90.3%) fell below 10% RRMSE, indicating that the hybrid estimator delivers consistently reliable sub-district welfare estimates for poverty targeting.

Keywords

Area Level; EBLUP; Fay Herriot–Random Forest; Lognormal transformation;

Full Text:

DOWNLOAD [PDF]

References

Breiman, L. (2001). Random Forests (R. E. Schapire, Ed.; Vol. 45). Kluwer Academic Publishers.

Bukhari, A. S., Notodiputro, K. A., Indahwati, & Fitrianto, A. (2025). A modified Fay-Herriot model with machine learning for small area estimation of per capita expenditure among the poor and agricultural households. BIO Web of Conferences, 186. https://doi.org/10.1051/bioconf/202518602012

Chandra, H., Aditya, K., & Kumar, S. (2018). Small-area estimation under a log-transformed area-level model. Journal of Statistical Theory and Practice, 12(3), 497–505. https://doi.org/10.1080/15598608.2017.1415174

Datta, G. S., & Ghosh, M. (2012). Small Area Shrinkage Estimation. Statistical Science, 27, 95–114.

Diffey, S. M., Smith, A. B., Welsh, A. H., & Cullis, B. R. (2017). A new REML (parameter expanded) EM algorithm for linear mixed models. Australian and New Zealand Journal of Statistics, 59(4), 433–448. https://doi.org/10.1111/anzs.12208

El Leithy, H. A., Abdel Wahed, Z. A., & Abdallah, M. S. (2016). On non-negative estimation of variance components in mixed linear models. Journal of Advanced Research, 7(1), 59–68. https://doi.org/10.1016/j.jare.2015.02.001

Esteban, M. D., Lombardía, M. J., López-Vizcaíno, E., Morales, D., & Pérez, A. (2020). Small area estimation of proportions under area-level compositional mixed models. Test, 29(3), 793–818. https://doi.org/10.1007/s11749-019-00688-w

Fauziah, G., Kurnia, A., & Djuraidah, A. (2025). The Empirical Best Linear Unbiased Prediction and The Emperical Best Predictor Unit-Level Approaches in Estimating Per Capita Expenditure at the Subdistrict Level. Scientific Journal of Informatics, 12(2), 283–294. https://doi.org/10.15294/sji.v12i2.25037

Fay, R. E., & Herriot, R. A. (1979). Estimates of Income for Small Places: An Application of James-Stein Procedures to Census Data. Journal of the American Statistical Association, 74(366a), 269–277. https://doi.org/10.1080/01621459.1979.10482505

Frink, N., & Schmid, T. (2024). Small area estimation with generalized random forests: Estimating poverty rates in Mexico. http://arxiv.org/abs/2406.03861

Haslett, S. J., & Welsh, A. H. (2019). EBLUPs : Empirical Best Linear Unbiased Predictors . In Wiley StatsRef: Statistics Reference Online (pp. 1–7). Wiley. https://doi.org/10.1002/9781118445112.stat08180

Hirose, M. Y. (2019). A class of general adjusted maximum likelihood methods for desirable mean squared error estimation of EBLUP under the Fay–Herriot small area model. Journal of Statistical Planning and Inference, 199, 302–310. https://doi.org/10.1016/j.jspi.2018.07.006

Jiang, J., & Rao, J. S. (2020). Annual Review of Statistics and Its Application Robust Small Area Estimation: An Overview. Downloaded from Www.Annualreviews.Org. Guest. https://doi.org/10.1146/annurev-statistics-031219

Krennmair, P., & Schmid, T. (2022). Flexible domain prediction using mixed effects random forests. Journal of the Royal Statistical Society. Series C: Applied Statistics, 71(5), 1865–1894. https://doi.org/10.1111/rssc.12600

Kurnia, A., Notodiputro, K. A., Saefuddin, A., & Mangku, I. W. (2009). Indonesia Case Small Area Estimation. ISSTEC 2009, 1–7.

Luis Avila-Valdez, J., Huerta, M., Leiva, V., & Trujillo, L. (2020). The Fay-Herriot Model in Small Area Estimation: EM Algorithm And Application To Official Data. REVSTAT - Statistical Journal, 18(5), 613–635. https://doi.org/10.57805/revstat.v18i5.323

Morales, D., Esteban, M. D., Pérez, A., & Hobza, T. (2021). Mean Squared Error of EBLUPs. In A Course on Small Area Estimation and Mixed Models (pp. 209–238). Springer. https://doi.org/10.1007/978-3-030-63757-6_9

Parker, P. A. (2024). Nonlinear Fay-Herriot Models for Small Area Estimation Using Random Weight Neural Networks. Journal of Official Statistics, 40(2), 317–332. https://doi.org/10.1177/0282423X241244671

Pfeffermann, D. (2013). New Important Developments in Small Area Estimation. Statistical Science, 28(1), 40 – 68. https://doi.org/10.1214/12-STS395

Rao, J. N. K., & Molina, I. (2015). Small Area Estimation (M. P. Couper, G. Kalton, J. N. K. Rao, Schwarz Norbert, & Skinner Christopher, Eds.; Second Edition). John Wiley &Sons, Inc. https://onlinelibrary.wiley.com/doi/

Rao, J. N. K., & Molina, I. (2016). Empirical bayes and hierarchical bayes estimation of poverty measures for small areas. In Analysis of Poverty Data by Small Area Estimation (pp. 315–324). wiley. https://doi.org/10.1002/9781118814963.ch17

Ratnaparkhi, M. (2014). Lognormal Distribution ‐ Basic. In Wiley StatsRef: Statistics Reference Online. Wiley. https://doi.org/10.1002/9781118445112.stat05889

Schonlau, M., & Zou, R. Y. (2020). The random forest algorithm for statistical learning. Stata Journal, 20(1), 3–29. https://doi.org/10.1177/1536867X20909688

Siems, I., & Münnich, R. (2025). Assessing the quality of generalized variance functions. AStA Wirtsch Sozialstat Arch, 19(3–4), 205–221. https://doi.org/https://doi.org/10.1007/s11943-025-00365-5

Slud, E. V, & Maiti, T. (2006). Mean-squared error estimation in transformed Fay-Herriot models. Journal of the Royal Statistical Society Series B: Statistical Methodology, 68(2), 239–257. https://doi.org/https://doi.org/10.1111/j.1467-9868.2006.00542.x

Tanur, E., Kurnia, A., Notodiputro, K. A., & Soleh, A. M. (2021). Log-normal small area estimation with measurement error- A n application for estimating household consumption expenditure. ICIC Express Letters, 15(5), 449–456. https://doi.org/10.24507/icicel.15.05.449

Ubaidillah, A., Notodiputro, K. A., Kurnia, A., & Mangku, I. W. (2019). Multivariate Fay-Herriot models for small area estimation with application to household consumption per capita expenditure in Indonesia. Journal of Applied Statistics, 46(15), 2845–2861. https://doi.org/10.1080/02664763.2019.1615420

Valliant, R. (1987). Generalized Variance Functions in Stratified Two-Stage Sampling. Journal of the American Statistical Association, 82(398), 499–508. https://doi.org/10.1080/01621459.1987.10478455

Viljanen, M., Meijerink, L., Zwakhals, L., & van de Kassteele, J. (2022). A machine learning approach to small area estimation: predicting the health, housing and well-being of the population of Netherlands. International Journal of Health Geographics, 21(1). https://doi.org/10.1186/s12942-022-00304-5

Wolter, K. M. (2007). Introduction to Variance Estimation (2nd ed.). Springer. https://doi.org/10.1007/978-0-387-35099-8

You, Y., & Hidiroglou, M. (2023). Application of Sampling Variance Smoothing Methods for Small Area Proportion Estimation. Journal of Official Statistics, 39(4), 571–590. https://doi.org/10.2478/jos-2023-0026

DOI: https://doi.org/10.31764/jtam.v10i4.39931