Una propuesta de aprendizaje automático para predecir la pobreza

Contenido principal del artículo

Martín Solís-Salazar
Julio Madrigal-Sanabria

Resumen

Debido a la alta tasa de errores de inclusión y exclusión de los métodos tradicionales (Proxy Mean Test) utilizados para la identificación de hogares en condición de pobreza y la selección de los beneficiarios de los programas de asistencia social, esta investigación analizó diferentes perspectivas para predecir hogares en condición de pobreza, utilizando un modelo de aprendizaje automático basado en XGBoost. Los modelos propuestos se compararon con métodos de referencia. Los datos utilizados fueron tomados de la encuesta de hogares del 2019 de Costa Rica. Los resultados mostraron que al menos uno de nuestros enfoques utilizando XGBoost dan el mejor balance entre el error de exclusión e inclusión. El mejor modelo se construyó utilizando XGBoost con un enfoque de clasificación.

Detalles del artículo

Cómo citar
Solís-Salazar, M., & Madrigal-Sanabria, J. (2022). Una propuesta de aprendizaje automático para predecir la pobreza . Revista Tecnología En Marcha, 35(4), Pág. 84–94. https://doi.org/10.18845/tm.v35i4.5766
Sección
Artículo científico

Citas

S , Kidd and E, Wylde, E, “Targeting the Poorest: An Assessment of the Proxy Means Test Methodology” Technical report, AusAID, Washington, DC, 2011

L, McBride and A, Nichols, "Retooling poverty targeting using out-of-sample validation and machine learning", The World Bank Economic Review, vol. 32, no. 3, pp. 531-550, 2018.

D, Budlender “Considerations in Using Proxy Means Tests In Eastern Caribbean States”., St.Lucia, 2016.

Available:https://pdfs.semanticscholar.org/efab/96de659b7208f41382341cf206ac25 9838c.pdf. [Accessed: March. 18, 2020].

F, Delgado-Jiménez, “Efectividad en la selección de beneficiarios de los programas avancemos y bienestar familiar”, Economía y Sociedad, vol. 22, no. 52, pp. 1-24, 2017

A, Bah, “Finding the Best Indicators to Identify the Poor”,Working Paper 01-2013, Jakarta, Indonesia: National Team for the Acceleration of Poverty Reduction (TNP2K), 2013

C ,Brown. M Ravallion and D, Van de Walle, “A Poor Means Test? Econometric Targeting in Africa”, Working Paper 22919, Massachusetts, EEUU: National Bureau of Economic Research, 2016.

S, Kidd. B. Gelders and D, Bailey-Athias, “Exclusion by design: an assessment of the effectiveness of the proxy means test poverty targeting mechanism”, ESS Working Paper No.56, Geneva: International Labour Office, 2017

S, Ashwini, et al, “A Proxy Means Test for Sri Lanka,. Working Paper 8605, 2018

D. S Mapa. & , M.L.F, Albis, “New Proxy means test (PMT) models: improving targeting of the poor for social protection”. In 12th National Convention on Statistics, Manila, Philippines, 2013.

R. K , Dewi and A, Suryahadi, “The implications of poverty dynamics for targeting the poor: simulations using Indonesian data”, Working paper, SMERU Research Institute, Indonesia, 2014

J, Hussein. O, Nazih, “Poverty level characterization via feature selection and machine learning”, In 27th Signal Processing and Communications Applications Conference (SIU), Siva, Turkey, 2019.

M, Pisacha, “Better Model Selection for poverty targeting through Machine Learning: A case Study in Thailand”, M.S. thesis, Thammasat University, Thailand, 2017

T, Pave and N, Stender, “Is random forest a superior methodology for predicting poverty? an empirical assessment”. Poverty & Public Policy, vol. 9, no. 1, pp. 118-133, 2017

T, Chen and C Guestrin, “Xgboost: A scalable tree boosting system”, In Proceedings of the 22nd ACM SIGKDD Conference on Knowledge Discovery and Data Mining, EEUU, San Francisco, CA, USA, 13 August 2016.

Guyon, et al, “Gene selection for cancer classification using support vector machines”. Machine learning, vol 46, no.1, pp. 389-422, 2002

H. Peng , F. Long , C. Ding, “Feature selection based on mutual information: cri- teria of max-dependency, max-relevance, and min-redundancy”, IEEE Trans. Pattern Anal. Mach. Intell. Vol. 27, pp. 1226–1238, 2005

S. García. J, Luengo. F. Herrera. “Tutorial on practical tips of the most influential data preprocessing algorithms in data mining”, Knowledge-Based Systems, vol. 98, pp. 1-29, 2016

O, Poursaeed. T, Matera, T. S, Belongie, “Vision-based real estate price estimation”. Machine Vision and Applications, vol. 29, no. 4, pp. 667-676, 2018

Y. Yoshimura et al, “Deep learning architect: classification for architectural design through the eye of artificial intelligence”. In International Conference on Computers in Urban Planning and Urban Management (pp. 249-265). Springer, Cham, 2019

R, Ngestrini, “Predicting Poverty of a Region from Satellite Imagery using CNNs”, M.S. thesis, Utrecht University, Utrecht, 2019

S.M, Pandey. T, Agarwal. N.C, Krishnan, “Multi-task deep learning for predicting poverty from satellite images”, In Thirty-Second AAAI Conference on Artificial Intelligence, New Orleans, USA, 2018

N. Jean, et al, “Combining satellite imagery and machine learning to predict poverty”. Science, vol. 353, pp. 790-794, 2016

V.H, Maluleke. S, Er. Q. R Williams, “Estimating poverty using aerial images: South African application”. Data Science and Applications, vol.1, no. 1, pp. 29-36, 2018.