RIDERSHIP PREDICTION SYSTEM FOR RAPID BUS KUANTAN AND PENANG USING MULTI-FEATURE ANALYSIS

Public transportation systems in Malaysia, particularly bus services in urban areas such as Kuantan and Penang, often struggle with inefficient scheduling and underutilized resources due to a lack of accurate demand forecasting tools. This study addresses the problem by developing a Ridership Prediction System for Rapid Bus Kuantan and Penang, employing a multi-feature analysis approach to predict bus ridership patterns based on various influencing factors, including weather conditions, weekends, and public holidays. The system aims to enhance public transportation efficiency through data-driven decision-making. A dataset covering ridership data from 2022 to 2024 was compiled from multiple sources, including Malaysia’s Official Open Data Portal, Open-Meteo, and Office Holidays. The study employs several machine learning models, including Multiple Linear Regression (MLR), Ridge Regression (RR), Support Vector Regression (SVR), Artificial Neural Networks (ANN), and Random Forest (RF), to analyze ridership trends and forecast demand. Evaluation results indicate that the Random Forest model outperforms other models, achieving the highest accuracy with an R-squared value of 0.9940, making it the optimal choice for prediction. A Graphical User Interface (GUI) is also developed to visualize ridership forecasts, enabling transit operators to make informed scheduling and resource allocation decisions. The system offers valuable insights for enhancing public transportation services and improving the passenger experience. Future improvements include integrating real-time Internet of Things (IoT)-based data collection and exploring advanced deep learning techniques to further refine prediction accuracy.