Application of survival analysis and machine learning models to age at first marriage among women in South Africa

Abstract

Understanding factors influencing the age at first marriage is crucial for addressing social issues, promoting gender equality, and ensuring women’s wellbeing. This research aims to identify key determinants of age at first marriage for South African women. The discrete survival tree approach is applied to analyze data and identify factors influencing women’s age at first marriage. Key individual variables, such as birth year, ethnicity, education level, age at first marriage, and province, are used in the analysis. The performance of this model is compared with that of Random Forests and Classification and Regression Trees using the C-index to determine the best-performing model.” All three models provided valuable insights, but Random Forest emerged as the most accurate age predictor at first marriage. Key determinants identified were province of residence, birth year, and educational level. These findings can contribute to policy-making aimed at improving the well-being of women in South Africa through targeted interventions.