Skip to main content

Research Repository

Advanced Search

Multinominial logistic regression modelling for perception evaluation of commuters to work.

Gupta, Kamini; Kumar, Ravindra; Tomar, S


Kamini Gupta

Ravindra Kumar

S Tomar


Commuting to work by bus is still one of the major transportation system in Delhi. Delhi government has introduced several reforms in services and it is found that the per capita trip rate (excluding walk trips) has increased from 0.72 in 1981 to 0.87 in 2001. It is estimated that per capita trip rate may reach to 1.2 by 2021 in Delhi. There is always a gap between satisfactions of user due to increase in trip rate and offered services by Delhi Transport Corporation. The major challenges to provide services are in terms of quality service indicators such as travel time, comfort, information system, accessibility, safety, different type of service offered, customer service, environmental impact etc. to improve overall quality by transport authority. To know the service performance of bus transportation one of the good method is User Perception Survey. The users’ perception about services offered by transit system plays a essential role in its success. It is therefore important to know the parameters which significantly influence user’s perception regarding bus services, knowing which one will impact more to improve the quality of transport services, as the public transportation improves the quality of life across the country by providing safe, economical and efficient services. In this paper research work is presented for evaluation of user perception of commuting passenger coming to CRRI Office in Delhi for official work. The age group of the passenger is considered adult those who are coming to office from different directions of Delhi. User Perception data has been collected by face to face interview. Respondents were asked to give their perception of the quality level across the different factors as well as satisfaction level toward their Bus services. A variable reduction technique (PCA) was used to identify influential variables & using these variables a Multinomial Logistic (MNL) Regression model (using STATA) was formed for Commuters satisfaction & their Comfort level. The model showed a higher degree of precision when compared with real-life data.

Presentation Conference Type Conference Paper (unpublished)
Conference Name Colloquium on Transportation Systems Engineering and Management
Start Date May 1, 2014
End Date May 1, 2014
Publication Date 2014
Deposit Date Jul 24, 2014
Peer Reviewed Not Peer Reviewed
Keywords bus user perception; mnl; official commute; transportation;
Public URL