Trend Analysis of Early Marriage Cases in South Sulawesi Using VAR for Stunting Solution

Abstract

This study uses a Vector Autoregressive (VAR) model to forecast early marriage cases across 23 regions in South Sulawesi. Monthly data from January 2017 to September 2024 were obtained from the provincial agency responsible for women’s empowerment and child protection and the Makassar Religious High Court. The analysis included stationarity testing, lag selection using the Akaike Information Criterion (AIC), parameter estimation, and evaluation using Mean Absolute Error (MAE). Among the candidate lags reported, VAR(2) was selected because it had the lowest AIC value, −69.241. Regional MAE values ranged from 0.666 to 4.650 cases. Forecasts for October 2024 to September 2025 showed differing regional patterns. Sengkang increased from 12 to 23 cases, Pangkajene from 2 to 7, and Pare-Pare from 4 to 6, whereas Barru and Masamba declined to 6 cases each. Makale and Selayar remained stable at 2 cases per month. These forecasts provide a basis for identifying regional priorities for early marriage prevention. Their relevance to stunting prevention is discussed in the context of maternal and child health; stunting itself was not included as a modeled variable.