Multivariate Adaptive Regression Splines for Contraceptive Use

Contraception is a critical public health issue, particularly in countries with high birth rates. In Indonesia, family planning has long been a priority, but challenges remain, especially in regions like West Sumatra, where socio-cultural factors influence contraceptive use. This study investigates the factors affecting contraceptive use in West Sumatra, focusing on socio-economic, cultural, and demographic variables. Traditional linear models often fail to capture the non-linear relationships and interactions between variables. To address this, the study employs the Multivariate Adaptive Regression Splines (MARS) method, a non-parametric regression technique that adapts to non-linearities and interactions. The model incorporates variables such as age, education, income, and access to family planning services, and identifies critical points (knots) where significant changes in contraceptive use occur. Results show that socio-economic factors, particularly education, income, and access to healthcare, significantly influence contraceptive use. The model demonstrates strong predictive performance, with an R² value of 0.98, indicating it explains 98% of the variability in contraceptive use patterns. By identifying key factors and their interactions, this study offers valuable insights for policymakers and public health officials, suggesting that improving access to services and addressing socio-cultural barriers can enhance family planning participation. The flexibility and predictive accuracy of the MARS model make it a valuable tool for evidence-based family planning interventions in West Sumatra and similar regions.