A Hybrid Machine Learning – Bayesian Spatial Survival Model of Early Sexual Initiation in Nigeria
International Journal of Development Mathematics (IJDM) · 12.06.2026 · 3 (2) · 333-350
DOI: 10.62054/ijdm/0302.21 ↗Kurz gefasst
Fachartikel in der Zeitschrift International Journal of Development Mathematics (IJDM), online erschienen am 12.06.2026. Verfasst von Paul Olopha. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).
Abstract (Originalsprache)
Early sexual initiation remains a major public health concern in Nigeria due to its links with sexually transmitted infections, unintended pregnancies, and unsafe abortions. This study investigated the determinants and spatial distribution of early sexual initiation using a hybrid framework combining machine learning and Bayesian spatial survival modelling. Data were obtained from the 2024 Nigeria Demographic and Health Survey, comprising 39,050 respondents. Time to first sexual intercourse was analysed within a survival framework while accounting for right censoring. A Random Survival Forest model was first used for variable selection, and the selected predictors were incorporated into a Bayesian structured additive spatial Weibull survival model estimated using Integrated Nested Laplace Approximation. Model performance was assessed using the concordance index, AIC, DIC, and WAIC. The hybrid model showed superior predictive performance, with higher time-dependent concordance indices (0.78–0.86) than conventional models. Higher educational attainment and household wealth significantly reduced the hazard of early sexual initiation, whereas urban residence increased risk. A nonlinear age effect showed sharply rising hazards during ages 15–18 years. Spatial analysis revealed higher risks in northern Nigeria and lower risks in southern regions. These findings underscore the need for targeted interventions addressing educational and socioeconomic inequalities in high-risk regions.
Zitieren
APA
Olopha, P. (2026). A Hybrid Machine Learning – Bayesian Spatial Survival Model of Early Sexual Initiation in Nigeria. International Journal of Development Mathematics (IJDM), 3(2), 333-350. https://doi.org/10.62054/ijdm/0302.21
Vancouver
Olopha P. A Hybrid Machine Learning – Bayesian Spatial Survival Model of Early Sexual Initiation in Nigeria. International Journal of Development Mathematics (IJDM). 2026;3(2):333-350. doi:10.62054/ijdm/0302.21
BibTeX
@article{Olopha2026,
author = {Olopha, Paul},
title = {A Hybrid Machine Learning – Bayesian Spatial Survival Model of Early Sexual Initiation in Nigeria},
journal = {International Journal of Development Mathematics (IJDM)},
year = {2026},
volume = {3},
number = {2},
pages = {333--350},
doi = {10.62054/ijdm/0302.21}
}
EUVIMED