Total Population Forecasting in Bangkok: A Comparative Study of Time Series and Machine Learning Models

This article conducts a comparative analysis of seven models to forecast Bangkok’s total population. The models evaluated include the Autoregressive Integrated Moving Average (ARIMA), Multilayer Perceptron (MLP), Support Vector Regression (SVR), Long Short-Term Memory (LSTM), and three hybrid models combining ARIMA with MLP, SVR, and LSTM. The study utilizes monthly population data from 2002 to 2022, sourced from the Registration Administration Office. The data is divided into a training set (2002–2021) and a test set (2022). Performance evaluation is based on Mean Squared Error (MSE), and Mean Absolute Percentage Error (MAPE). Among the models, the LSTM model exhibited superior performance, with the lowest MSE (14,821.75), and MAPE (0.269), outperforming both traditional models and hybrid approaches. These results suggest that the LSTM model is the most effective for accurate population forecasting in Bangkok, providing valuable insights for urban planning and resource allocation.