World Journal of Engineering Research and Technology

Anika Nowshin Mowrin* and D. M. Ghius Malik

ABSTRACT

Traffic congestions is a common problem in urban transportation system of major cities in Bangladesh. In recent time, the possibility of introducing commuter train services in urban cities are explored. It has been assess that the service of commuter trains can be a vital solution of traffic congestion. This paper deals with the assessment of service quality of commuter train based on users perception data collected from different cities of Bangladesh. For this purpose, a questionnaire survey have been carried out among 802 respondents who traveled in commuter train as their daily mode of transport. In order to evaluatethe performance of existing commuter train service in the transportation system, two models have been used. These are (i) Adaptive Neuro Fuzzy Inference System (ANFIS) (ii) Probabilistic Neural Network (PNN). Accuracy of ANFIS and PNN have been evaluated using confusion matrix and root mean square error (RMSE). The efficacy of ANFIS and PNN found to be 61.50% and 67.40% respectively in training period. Moreover, in testing the models, the accuracy have been found as 47.80% and 44.10% respectively. Comparison shows that ANFIS is most suitable for commuter train service quality prediction than that of PNN. Finally, a stepwise approach was followed for ranking the commuter train service quality attributes and the results were compared with that of the public opinions. From the results, it is found that 'Bogie condition', 'Cleanliness', 'Behavior of staff', 'Toilet facility' are the most significant attributes. It is hoped that the paper will be helpful for the decision makers for improving the service quality of urban transportation system.

[Full Text Article] [Download Certificate]