Predicting Birth Rates Using Random Forest, Long Short-Term Memory and Exponential Smoothing: A Study on Kotawaringin Barat Regency, Central Kalimantan, Indonesia

Predicting birth rates accurately is essential for effective population planning. Traditional time series methods like Exponential Smoothing have been applied to birth rate forecasting but often fail to achieve high accuracy. This work suggests a novel approach using advanced machine learning, notably Random Forest (RF) and Long Short-Term Memory (LSTM), to predict birth rates in Kotawaringin Barat Regency, Central Kalimantan Province, Indonesia, using daily birth records from 2015 to 2023. The contribution of this study lies in evaluating the performance of these methods against Exponential Smoothing to determine which model is the most successful for this context. From the evaluation, this research predicts birth rates with Exponential Smoothing and LSTM models resulting in MAPE values of 0.32 and 0.38 and the lowest error, achieving 0.09 MAPE using the Random Forest method.