SISTEM DETEKSI CERDAS BERBASIS METODE DEMPSTER-SHAFER UNTUK MENDIAGNOSIS KERUSAKAN PADA SEPEDA MOTOR INJEKSI Artificial Intelligence, Dempster-Shafer, Expert System, Fuel Injection System and Motorcycle

Abstract

Injection motorcycles are an innovation in two-wheeled vehicles that integrate electronic technology to enhance efficiency and performance. However, many users face difficulties in detecting electrical system failures in these motorcycles, primarily due to a lack of technical knowledge. This study aims to develop a web-based expert system capable of detecting issues in motorcycle injection components, specifically for injection motorcycles of matic and underbone types from 2013 to 2019. The system utilizes the Dempster-Shafer method to handle uncertainty in the diagnostic process. The Dempster-Shafer theory provides reasoning based on belief functions and plausible reasoning, which is used to combine separate pieces of information (evidence) to calculate the likelihood of an event, resulting in more accurate outcomes. The input consists of data from 15 symptoms of injection component failures and 7 types of injection component failures. The output of this web-based expert system is the diagnosis of the detected failures in the motorcycle injection components and the solutions for these failures. The results show that the developed expert system can improve efficiency in detecting failures and provide more accurate solutions for users with an accuracy rate of 90% based on the expert system's results and the direct testing on 10 injection motorcycle units.