A comparative study of machine learning method and spatial filtered negative binomial model on handling spatial correlation when predicting the daily demand of shared bike trips

Lei Gong; Qin Luo

doi:10.1117/12.2619715

22 December 2021 A comparative study of machine learning method and spatial filtered negative binomial model on handling spatial correlation when predicting the daily demand of shared bike trips

Lei Gong, Qin Luo

Author Affiliations +

Proceedings Volume 12058, Fifth International Conference on Traffic Engineering and Transportation System (ICTETS 2021); 120581S (2021) https://doi.org/10.1117/12.2619715
Event: 5th International Conference on Traffic Engineering and Transportation System (ICTETS 2021), 2021, Chongqing, China

Abstract

Spatial correlation is an important issue for the accurate prediction of demand of shared bikes in the service area, as it exists and has an unignorable effect on the results when making demand forecast based on a traffic zone such as tract. This study makes a comparative study by predicting the daily demand of shared bikes in each tract between a negative binomial regression model including spatial filters and a random forest model. Several performance indexes including MAE, RMSE and MAPE are used for the comparison of predicting the daily demand with a consideration of spatial correlation. The results show that as a widely used machine learning method, random forest model shows a better performance than the spatial filtered negative binomial model, which means the former includes the spatial pattern in its algorithm when learning the general pattern in the data.

Citation Download Citation

Lei Gong and Qin Luo "A comparative study of machine learning method and spatial filtered negative binomial model on handling spatial correlation when predicting the daily demand of shared bike trips", Proc. SPIE 12058, Fifth International Conference on Traffic Engineering and Transportation System (ICTETS 2021), 120581S (22 December 2021); https://doi.org/10.1117/12.2619715

ACCESS THE FULL ARTICLE

INSTITUTIONAL
Select your institution to access the SPIE Digital Library.

SELECT YOUR INSTITUTION

PERSONAL
Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.

PERSONAL SIGN IN

No SPIE Account? Create one

PURCHASE THIS CONTENT

SUBSCRIBE TO DIGITAL LIBRARY

50 downloads per 1-year subscription

Members: $195

Non-members: $335 ADD TO CART

25 downloads per 1 - year subscription

Members: $145

Non-members: $250 ADD TO CART

PURCHASE SINGLE ARTICLE

Includes PDF, HTML & Video, when available