Development of a Provincial Flood Prediction Model for Thailand in 2025 Using Machine Learning Techniques

Abstract

The study aims to develop a provincial flood prediction model for Thailand in 2025, employing projected rainfall data and flood statistics from 2021 to 2023. The model is essential for catastrophe preparedness and future risk management strategies. This study utilizes statistical analysis alongside machine learning methodologies employing the Random Forest algorithm to develop a highly accurate model for forecasting flood-prone regions. The analytical procedure include data preprocessing, including the identification of missing values, the management of data redundancy, and the preparation of the dataset prior to model training and testing. The data include monthly precipitation forecasts for the forthcoming six months and yearly flood statistics. The findings of this study can be utilized to enhance decision-making and risk management regarding floods by pertinent agencies. The findings establish a basis for the future advancement of more efficient predictive algorithms.