BRP-Net: A discrete-aware network based on attention mechanisms and LSTM for birth rate prediction in prefecture-level cities

Mingfu Xue,Junyu Zhu,Rusheng Wu,Xiayiwei Zhang,Yuan Chen

DOI: https://doi.org/10.1371/journal.pone.0307721

IF: 3.7

2024-09-12

PLoS ONE

Abstract:The continuous decline in the birth rate can lead to a series of social and economic problems. Accurately predicting the birth rate of a region will help national and local governments to formulate more scientifically sound development policies. This paper proposes a discrete-aware model BRP-Net based on attention mechanism and LSTM, for effectively predicting the birth rate of prefecture-level cities. BRP-Net is trained using multiple variables related to comprehensive development of prefecture-level cities, covering factors such as economy, education and population structure that can influence the birth rate. Additionally, the comprehensive data of China's prefecture-level cities exhibits strong spatiotemporal specificity. Our model leverages the advantages of attention mechanism to identify the feature correlation and temporal relationships of these multi-variable time series input data. Extensive experimental results demonstrate that the proposed BRP-Net has higher accuracy and better generalization performance compared to other mainstream methods, while being able to adapt to the spatiotemporal specificity of variables between prefecture-level cities. Using BRP-Net to achieve precise and robust prediction estimates of the birth rate in prefecture-level cities can provide more effective decision-making references for local governments to formulate more accurate and reasonable fertility encouragement policies.