Predicting Fertility Determinants in Bangladesh Using Machine Learning Models
- 1Department of Statistics, Mawlana Bhashani Science and Technology University, Bangladesh
- 2Department of Statistics, Shahjalal University of Science and Technology, Bangladesh
Abstract
Background: Fertility is a key demographic indicator that influences population dynamics, economic growth, social welfare, and public policy. This study aims to fill a gap in existing research by applying advanced machine learning (ML) methods to predict fertility outcomes and identify key fertility determinants in Bangladesh using data from the Bangladesh Demographic Health Survey (BDHS) 2017–18.
Methods: The dataset comprised 20,127 women, reflecting the fertility landscape in Bangladesh. We employed both conventional statistical methods and various machine learning models including Random Forest (RF), Decision Tree (DT), K-Nearest Neighbors (KNN), Logistic Regression (LR), Support Vector Machine (SVM), XGBoost, LightGBM and Neural Networks (NN). The performance of the models was evaluated based on metrics such as accuracy, precision, recall, F1 score, Area Under the Receiver Operating Characteristic Curve (AUROC) and alongside K-fold cross-validation. .
Results: The SVM model emerged as the most effective, with macro average recall (89%), precision (89%) and F1 score (89%) as well as weighted average recall (96%), precision (95%), F1 score (95%) and accuracy was 95.5%. The results revealed that access to mass media, father’s education level, mother’s highest educational attainment, number of household members, maternal Body Mass Index (BMI), number of living children and child mortality (i.e., incidents of sons or daughters dying) were the most influential factors of fertility in Bangladesh.
Conclusions: Our findings demonstrate that machine learning models, particularly SVM, provide superior predictive power compared to traditional statistical methods in identifying key fertility determinants. These insights contribute to a deeper understanding of fertility dynamics in Bangladesh and suggest that targeted interventions—focusing on media access, parental education, and healthcare improvements could significantly influence fertility rates. Future research should explore the use of ensemble models and longitudinal data to enhance predictive accuracy and capture temporal changes in fertility determinants.
Related
- Machine learning analysis of the association between fertility and under five mortality in Bangladesh using BDHS 2022 data
- Trends, patterns, determinants and socio-economic inequality of high-risk fertility behavior (HRFB) among Bangladeshi women: Evidence from Bangladesh Demographic and Health Surveys
- Double machine learning analysis identifies heterogeneous patterns in Ethiopian fertility intention: cross-sectional study
- Explainable machine learning and deep learning models for predicting in-hospital mortality in burn patients
- Predicting MCI-to-Alzheimer’s Conversion: A Systematic Benchmark of Classical, Machine Learning, and Deep Survival Models
BibTeX
@article{naznin2026predicting,
title = {{Predicting Fertility Determinants in Bangladesh Using Machine Learning Models}},
author = {Shayla Naznin and Dr. Md Jamal Uddin and Dr. Ahmad Kabir},
journal = {Asia Pacific Journal of Health Management},
year = {2026},
month = nov,
doi = {10.24083/apjhm.v21i2.5407},
url = {https://doi.org/10.24083/apjhm.v21i2.5407},
note = {CC BY-NC},
}
RIS
TY - JOUR
TI - Predicting Fertility Determinants in Bangladesh Using Machine Learning Models
AU - Shayla Naznin
AU - Dr. Md Jamal Uddin
AU - Dr. Ahmad Kabir
PY - 2026
DA - 2026/11/01
JO - Asia Pacific Journal of Health Management
DO - 10.24083/apjhm.v21i2.5407
UR - https://doi.org/10.24083/apjhm.v21i2.5407
AB - Background: Fertility is a key demographic indicator that influences population dynamics, economic growth, social welfare, and public policy. This study aims to fill a gap in existing research by applying advanced machine learning (ML) methods to predict fertility outcomes and identify key fertility determinants in Bangladesh using data from the Bangladesh Demographic Health Survey (BDHS) 2017–18. Methods: The dataset comprised 20,127 women, reflecting the fertility landscape in Bangladesh. We employed both conventional statistical methods and various machine learning models including Random Forest (RF), Decision Tree (DT), K-Nearest Neighbors (KNN), Logistic Regression (LR), Support Vector Machine (SVM), XGBoost, LightGBM and Neural Networks (NN). The performance of the models was evaluated based on metrics such as accuracy, precision, recall, F1 score, Area Under the Receiver Operating Characteristic Curve (AUROC) and alongside K-fold cross-validation. . Results: The SVM model emerged as the most effective, with macro average recall (89%), precision (89%) and F1 score (89%) as well as weighted average recall (96%), precision (95%), F1 score (95%) and accuracy was 95.5%. The results revealed that access to mass media, father’s education level, mother’s highest educational attainment, number of household members, maternal Body Mass Index (BMI), number of living children and child mortality (i.e., incidents of sons or daughters dying) were the most influential factors of fertility in Bangladesh. Conclusions: Our findings demonstrate that machine learning models, particularly SVM, provide superior predictive power compared to traditional statistical methods in identifying key fertility determinants. These insights contribute to a deeper understanding of fertility dynamics in Bangladesh and suggest that targeted interventions—focusing on media access, parental education, and healthcare improvements could significantly influence fertility rates. Future research should explore the use of ensemble models and longitudinal data to enhance predictive accuracy and capture temporal changes in fertility determinants.
SN - 1833-3818
SN - 2204-3136
ER -