ANALISIS KERAWANAN MULTI-BAHAYA BANJIR DAN LONGSOR KOTA BANDUNG MENGGUNAKAN INTEGRASI ANALISIS MULTI-KRITERIA DAN MACHINE LEARNING

• Main Author: Putri, Ramadani
• Format: Final Project
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/65210
• Institution: Institut Teknologi Bandung

As a tropical climate region in the world, Indonesia has a high level of vulnerability to hydrometeorological disasters. Risk and disaster are two things that cannot be separated and cannot be eliminated, but they can be minimized. One of the efforts that can be done is to analyze the possible dangers that can occur. The spatial modeling of the hazard of an area is studied from only one type of disaster. Data from the National Disaster Management Agency (BNPB), Bandung City found a "medium" risk class for 2020. The topographical character of Bandung City as a city below between mountains with a lack of air catchment areas due to rapid population growth and human activities that cause flooding risk. This study tries to model floods and landslides as multi-hazard disasters in Bandung City. The modeling is done using the Multi-Criteria Analysis (MCA) and Machine Learning methods with three approaches: support vector machine (SVM), Classification and Regression Trees (CART), and Random Forest (RF). The flood and landslide inventory data were randomly divided into (70%) sample data and (30%) data validation. The resulting model has an overall accuracy of 0.71 for landslide hazard and 0.866 for flood hazard.