Monthly river flow forecasting in Kelantan with ARIMA and deep learning LSTM

Teoh, Xu Xian (2025) Monthly river flow forecasting in Kelantan with ARIMA and deep learning LSTM. Final Year Project, UTAR.

PDF Download (2909Kb)

Abstract

This project aims to develop a web-based river flow forecasting system tailored to Malaysian rivers by integrating two prominent time series forecasting models: ARIMA and Long Short-Term Memory (LSTM). The system focuses on Sungai Kelantan and Sungai Sokor, leveraging daily river discharge data sourced from the Department of Irrigation and Drainage (DID) Malaysia. The core objective is to deliver accurate monthly forecasts through a user-friendly interface powered by Streamlit. The methodology follows the CRISP-DM framework, including systematic data preprocessing, model training, and evaluation using Root Mean Square Error (RMSE), Mean Absolute Error (MAE), and classification metrics. Forecast accuracy, especially for extreme flow conditions, is validated through comparative performance analysis. The final product allows real-time river flow forecasting with interactive model selection and visualization, contributing to improved decision-making for flood preparedness and water resource management in Malaysia.

Actions (login required)