91ÊÓƵ

Predictive Modelling in Flood Area Using Artificial Intelligence and Machine Learning Methods: Study Case West Java Province

Sri Supatmi, Mia Firiawati, Hanhan Maulana, Rongtao Hou, Shounan Hou

Abstract

West Java Province is one of the provinces in Indonesia that was impacted by flood disasters during the rainy season. An experimental investigation was conducted to explore the fundamental, predictive models among artificial intelligence and machine learning, and the proposed flood predictive model is called a hybrid model, combining two models: artificial intelligence and machine learning. The study aims to find which models give the best performance for flood susceptibility prediction. The comparison of Artificial Intelligence (AI), Machine Learning, and proposed hybrid predictive models is an essential step in accurately predicting flood event vulnerability. This research employed AI (Artificial Neural Network), machine learning (SVM), and a proposed predictive model combining Artificial Neural Network and SVM to predict the time series data for 31 districts in Bandung, West Java Province, Indonesia, between 2012 and 2023. Our findings revealed that the proposed model exhibits a flood vulnerability predicting accuracy exceeding 98.5%, with the model errors being the lowest compared to existing models. These results indicate the efficacy of the proposed model in predicting flood vulnerability, which can be a valuable tool for businesses and policymakers operating in regions such as West Java. Our study contributes to the growing literature on flood vulnerability prediction and highlights the importance of utilizing advanced modeling techniques to improve the accuracy of predictions.

keyword

Flood, Artificial Intelligence, Machine learning, predictive model, West Java

Area

Data Mining and Optimization