Rekonstruksi Multidimensional Poverty Index dengan XGBoost dan Tipologi Deprivasi: Label Klaster Redundan sebagai Prediktor pada Data Global MPI 2025
Kata Kunci:
Multidimensional Poverty Index, XGBoost, K-Means, Rekonstruksi, Indeks, GroupKFoldAbstrak
Skor Multidimensional Poverty Index (MPI) menyembunyikan komposisi deprivasi, sehingga negara dengan skor berdekatan dapat memerlukan intervensi berbeda. Sebagian penerapan machine learning pada data MPI melaporkan akurasi tinggi tanpa membedakan rekonstruksi rumus indeks dari prediksi variabel di luar indeks, dan tanpa menguji apakah label klaster menambah informasi. Penelitian ini memisahkan kedua hal itu pada 186 observasi negara-tahun dari 67 negara (Global MPI 2025, survei 2001–2023). XGBoost merekonstruksi MPI dari sepuluh indikator deprivasi dengan validasi GroupKFold lima lipatan menurut negara, agar survei berulang satu negara tidak berada di kedua sisi lipatan, sedangkan K-Means membentuk tipologi deprivasi empat klaster. Rekonstruksi mencapai R² 0,988, yang menegaskan konsistensi konstruk, bukan kemampuan meramal. Identitas klaster tidak menambah akurasi (selisih R² 0,0001) dan hasil nol itu bertahan setelah label dicocokkan antarlipatan serta dikodekan ulang, sehingga klaster redundan sebagai prediktor meskipun tetap berguna sebagai tipologi pemantauan: 25 dari 67 negara berpindah klaster antarsurvei. Bahan bakar memasak paling banyak dipakai oleh model rekonstruksi, namun korelasi antarindikator yang tinggi membatasi tafsir peringkatnya. Prioritas program sebaiknya mengikuti posisi deprivasi pada survei terakhir, bukan peringkat maupun skor yang dihasilkan model.
Referensi
Aas, K., Jullum, M., & Løland, A. (2021). Explaining individual predictions when features are dependent: More accurate approximations to Shapley values. Artificial Intelligence, 298, 103502. https://doi.org/10.1016/j.artint.2021.103502
Alkire, S., Kanagaratnam, U., & Suppa, N. (2024). The global Multidimensional Poverty Index (MPI) 2024: Country results and methodological note. Oxford Poverty and Human Development Initiative, University of Oxford. https://ophi.org.uk/publications/MN-58
Alkire, S., Mishra, R., Selden, L., & Suppa, N. (2025a). A methodological note on the global Multidimensional Poverty Index (MPI) 2025: Changes over time results for 88 countries. Oxford Poverty and Human Development Initiative, University of Oxford. https://ophi.org.uk/publications/MN-63
Alkire, S., Mishra, R., Selden, L., & Suppa, N. (2025b). The global Multidimensional Poverty Index (MPI) 2025: Country results and methodological note. Oxford Poverty and Human Development Initiative, University of Oxford. https://ophi.org.uk/publications/MN-61
Alsharkawi, A., Al-Fetyani, M., Dawas, M., Saadeh, H., & Alyaman, M. (2021). Poverty classification using machine learning: The case of Jordan. Sustainability, 13(3), 1412. https://doi.org/10.3390/su13031412
Browne, C., Matteson, D. S., McBride, L., Hu, L., Liu, Y., Sun, Y., Wen, J., & Barrett, C. B. (2021). Multivariate random forest prediction of poverty and malnutrition prevalence. PLoS ONE, 16(9), e0255519. https://doi.org/10.1371/journal.pone.0255519
Cerqua, A., Letta, M., & Pinto, G. (2025). On the (mis)use of machine learning with panel data. Oxford Bulletin of Economics and Statistics, 88(3), 506–518. https://doi.org/10.1111/obes.70019
Delahoz-Domínguez, E., Mendoza-Mendoza, A., & Visbal-Cadavid, D. (2025). Clustering of countries through UMAP and K-means: A multidimensional analysis of development, governance, and logistics. Logistics, 9(3), 108. https://doi.org/10.3390/logistics9030108
Hooker, G., Mentch, L., & Zhou, S. (2021). Unrestricted permutation forces extrapolation: Variable importance requires at least one more model, or there is no free variable importance. Statistics and Computing, 31(6), 82. https://doi.org/10.1007/s11222-021-10057-z
Hu, S., Ge, Y., Liu, M., Ren, Z., & Zhang, X. (2022). Village-level poverty identification using machine learning, high-resolution images, and geospatial data. International Journal of Applied Earth Observation and Geoinformation, 107, 102694. https://doi.org/10.1016/j.jag.2022.102694
Indah, Y. M., Aristawidya, R., Fitrianto, A., Erfiani, & Jumansyah, L. M. R. D. (2025). Comparison of Random Forest, XGBoost, and LightGBM methods for the Human Development Index classification. Jambura Journal of Mathematics, 7(1). https://doi.org/10.37905/jjom.v7i1.28290
Kumar, S., Chakraverty, S., & Sethi, N. (2023). Multidimensional poverty: CMPI development, spatial analysis and clustering. Social Indicators Research, 169, 647–670. https://doi.org/10.1007/s11205-023-03181-y
Li, Q., Yu, S., Échevin, D., & Fan, M. (2022). Is poverty predictable with machine learning? A study of DHS data from Kyrgyzstan. Socio-Economic Planning Sciences, 81, 101195. https://doi.org/10.1016/j.seps.2021.101195
Loecher, M. (2024). Debiasing SHAP scores in random forests. AStA Advances in Statistical Analysis, 108(2), 427–440. https://doi.org/10.1007/s10182-023-00479-7
Lundberg, S. M., Erion, G., Chen, H., DeGrave, A., Prutkin, J. M., Nair, B., Katz, R., Himmelfarb, J., Bansal, N., & Lee, S.-I. (2020). From local explanations to global understanding with explainable AI for trees. Nature Machine Intelligence, 2(1), 56–67. https://doi.org/10.1038/s42256-019-0138-9
Lyons, A. C., Montoya Castano, A., Kass-Hanna, J., Zhang, Y., & Soliman, A. (2025). A machine learning approach to assessing multidimensional poverty and targeting assistance among forcibly displaced populations. World Development, 192, 107013. https://doi.org/10.1016/j.worlddev.2025.107013
Meyer, H., Reudenbach, C., Wöllauer, S., & Nauss, T. (2019). Importance of spatial predictor variable selection in machine learning applications: Moving from data reproduction to spatial prediction. Ecological Modelling, 411, 108815. https://doi.org/10.1016/j.ecolmodel.2019.108815
Molnar, C. (2022). Interpretable machine learning: A guide for making black box models explainable (2nd ed.). https://christophm.github.io/interpretable-ml-book/
Muñetón-Santa, G., & Manrique-Ruiz, L. C. (2023). Predicting multidimensional poverty with machine learning algorithms: An open data source approach using spatial data. Social Sciences, 12(5), 296. https://doi.org/10.3390/socsci12050296
Olearo, L., D’Adda, F., Messina, E., Cremaschi, M., Bandini, S., & Gasparini, F. (2024). Facing multidimensional poverty in older adults: An artificial intelligence approach that reveals the variable relevance. Intelligenza Artificiale, 18(1), 51–65. https://doi.org/10.3233/IA-240027
Onsay, E. A., Alinsunurin, J., & Rabajante, J. F. (2025). Optimizing machine learning algorithms for multidimensional poverty prediction in the Philippines. SN Business & Economics, 5, 159. https://doi.org/10.1007/s43546-025-00922-8
Puttanapong, N., Martinez Jr., A., Bulan, J. A. N., Addawe, M., Durante, R. L., & Martillan, M. (2022). Predicting poverty using geospatial data in Thailand. ISPRS International Journal of Geo-Information, 11(5), 293. https://doi.org/10.3390/ijgi11050293
Rousseeuw, P. J. (1987). Silhouettes: A graphical aid to the interpretation and validation of cluster analysis. Journal of Computational and Applied Mathematics, 20, 53–65. https://doi.org/10.1016/0377-0427(87)90125-7
Sari, D. P., Rizqi, T. A., Fitrianto, A., Erfiani, & Jumansyah, L. M. R. D. (2024). Perbandingan metode K-Means dan OPTICS dalam penggerombolan kemiskinan multidimensi di Indonesia. Kubik: Jurnal Publikasi Ilmiah Matematika, 9(2). https://garuda.kemdiktisaintek.go.id/documents/detail/4840389
Satapathy, S. K., Saravanan, S., Mishra, S., & Mohanty, S. N. (2023). A comparative analysis of multidimensional COVID-19 poverty determinants: An observational machine learning approach. New Generation Computing, 41(1), 155–184. https://doi.org/10.1007/s00354-023-00203-8
Shwartz-Ziv, R., & Armon, A. (2022). Tabular data: Deep learning is not all you need. Information Fusion, 81, 84–90. https://doi.org/10.1016/j.inffus.2021.11.011
Takefuji, Y. (2025). Beyond XGBoost and SHAP: Unveiling true feature importance. Journal of Hazardous Materials, 488, 137382. https://doi.org/10.1016/j.jhazmat.2025.137382
UNDP & OPHI. (2023). Global Multidimensional Poverty Index 2023: Unstacking global poverty: Data for high impact action. United Nations Development Programme and Oxford Poverty and Human Development Initiative. https://hdr.undp.org/content/2023-global-multidimensional-poverty-index-mpi
UNDP & OPHI. (2024). Global Multidimensional Poverty Index 2024: Poverty amid conflict. United Nations Development Programme and Oxford Poverty and Human Development Initiative. https://hdr.undp.org/content/2024-global-multidimensional-poverty-index-mpi
UNDP & OPHI. (2025). Global Multidimensional Poverty Index 2025: Overlapping hardships: Poverty and climate hazards. United Nations Development Programme and Oxford Poverty and Human Development Initiative. https://hdr.undp.org/content/2025-global-multidimensional-poverty-index-mpi
Usmanova, A., Aziz, A., Rakhmonov, D., & Osamy, W. (2022). Utilities of artificial intelligence in poverty prediction: A review. Sustainability, 14(21), 14238. https://doi.org/10.3390/su142114238
World Bank. (2022). Poverty and shared prosperity 2022: Correcting course. World Bank. https://doi.org/10.1596/978-1-4648-1893-6
Xu, J., Song, J., Li, B., Liu, D., & Cao, X. (2021). Combining night time lights in prediction of poverty incidence at the county level. Applied Geography, 135, 102552. https://doi.org/10.1016/j.apgeog.2021.102552
Yang, J., Jin, J., Qiu, X., & Liu, D. (2025). Does the cooking fuel transition lead to unexpected energy poverty? Evidence from the less-developed Tibetan region of China. Energy Policy, 198, 114471. https://doi.org/10.1016/j.enpol.2025.114471
Zamzuri, M. H. A., Sofian, N., & Hassan, R. (2023). The forecasting of poverty using the ensemble learning classification methods. International Journal on Perceptive and Cognitive Computing, 9(1), 24–32. https://doi.org/10.31436/ijpcc.v9i1.326
Zhang, W., Lei, T., Gong, Y., Zhang, J., & Wu, Y. (2022). Using explainable artificial intelligence to identify key characteristics of deep poverty for each household. Sustainability, 14(16), 9872. https://doi.org/10.3390/su14169872
