Real-time flood forecasting in Amo Chhu using machine learning model and internet of things

Abstract

The proposed research aims to predict flood events in Amo Chhu river of Bhutan using machine learning and Internet of Things. The Amo Chhu River has been posing threat to the residents along river Amo Chhu, and this research seeks to improve flood forecasting to mitigate the impact of natural disasters on vulnerable communities. The research utilizes three machine learning models, including Random Forest, Gradient boosting, and Support Vector Classifier, in combination with IoT devices. The data required to train the machine learning models was obtained from National Centre for Hydrology and Meteorology (NCHM) ranged from 1996 to 2022. The current work evaluates the efficiency of these models in predicting flood events in Amo Chhu River. The proposed work found that the Gradient Boosting Classifier model delivered outstanding performance, achieving an accuracy of 99.8%, followed by Support Vector Classifier model with an accuracy of 98.3%. The result from Random Forest model showed promising outcome, with accuracy of 97.9%. The inclusion of real-time data from IoT devices significantly improved the effectiveness of the forecasting system thus indicating potential of machine learning models and IoT devices in improving flood forecasting and mitigating the impact of natural disasters on vulnerable communities.

Keywords

• Machine learning
• internet of things
• flood
• prediction
• random forest
• support vector classifier