Public transportation systems, particularly bus services, are essential for urban mobility, significantly influencing environmental sustainability and easing traffic congestion. This study explores the impact of weather conditions, special events, and holidays on bus ridership, using data from Kuala Lumpur's Rapid Bus KL service. A range of machine learning models—Linear Regression, ARIMA, Random Forest, and XGBoost—were applied to predict ridership and identify key influencing factors. The study found that advanced models, especially Random Forest achieved the highest performance, with an value of 0.9656, demonstrating its ability to explain 96.56% of the variance in ridership. Key features, such as weekdays versus weekends, temperature, and festivals, were identified as significant predictors of ridership patterns. The findings highlight the importance of incorporating these variables into operational planning to improve public transit service delivery and meet fluctuating demand, especially during peak times. The study concludes with suggestions for further research into enhancing prediction accuracy through refined model development and data integration.
Public Transport Ridership Forecasting Using Machine Learning: A Case Study of Rapid Bus KL in Malaysia
2024-12-20
547113 byte
Conference paper
Electronic Resource
English
Forecasting Passenger Ridership
British Library Conference Proceedings | 1998
|Transit Ridership Forecasting Using a GIS
British Library Conference Proceedings | 1998
|A Commuter Rail Ridership Forecasting Case Study from Southern California
British Library Conference Proceedings | 2007
|Forecasting bus ridership using a “Blended Approach”
Online Contents | 2019
|