A platform for research: civil engineering, architecture and urbanism
A platform for research: civil engineering, architecture and urbanism
Improving ridership by predicting train occupancy levels
With the frequent global breakouts of infectious diseases such as Covid-19 and the likes, passengers feel unsafe traveling in crowded trains. The reluctance to share public transport with others due to the risk of disease transmission may lower the ridership as well as decrease the comfort level of passengers. Providing them with future crowdedness levels may allow them to plan accordingly, hence regaining the lost confidence and improving their patronage. This study explores the less frequently investigated relationship among occupancy levels at a particular station over several train runs, to predict the future occupancy level with a delay of one run (day). Tackling the issue as a classification problem rather than a regression problem, train occupancy data, station data, and weather data are merged to develop the final dataset. Training data is stepwise increased from 1 month to 3 months. Similarly, 1–5 days of known occupancy levels are added to each data instance. Among the three classifiers used, XGBoost provides the best results. Some practical challenges to occupancy level prediction are also discussed at the end.
Improving ridership by predicting train occupancy levels
With the frequent global breakouts of infectious diseases such as Covid-19 and the likes, passengers feel unsafe traveling in crowded trains. The reluctance to share public transport with others due to the risk of disease transmission may lower the ridership as well as decrease the comfort level of passengers. Providing them with future crowdedness levels may allow them to plan accordingly, hence regaining the lost confidence and improving their patronage. This study explores the less frequently investigated relationship among occupancy levels at a particular station over several train runs, to predict the future occupancy level with a delay of one run (day). Tackling the issue as a classification problem rather than a regression problem, train occupancy data, station data, and weather data are merged to develop the final dataset. Training data is stepwise increased from 1 month to 3 months. Similarly, 1–5 days of known occupancy levels are added to each data instance. Among the three classifiers used, XGBoost provides the best results. Some practical challenges to occupancy level prediction are also discussed at the end.
Improving ridership by predicting train occupancy levels
Muhammad Awais Shafique (author)
2024
Article (Journal)
Electronic Resource
Unknown
Metadata by DOAJ is licensed under CC BY-SA 1.0
Time-Expanded Network Model of Train-Level Subway Ridership Flows Using Actual Train Movement Data
| British Library Online Contents | 2016
Ridership and cost on the Long Beach-Los Angeles Blue Line Train
| Elsevier | 1992
Ridership and cost on the Long Beach-Los Angeles Blue Line Train
| Online Contents | 1993
Ridership and cost on the Long Beach-Los Angeles Blue Line Train
| British Library Online Contents | 1993
How to Increase Rail Ridership in Maryland: Direct Ridership Models for Policy Guidance
| Online Contents | 2016