Train Service Planning Prototype Model For Thailand Hi-Speed Train

Abstract

Due to the efficiency of the transportation system is very important for a developing country’s economy. Good management transportation services will help agencies provide sustainability services and for users satisfied. This research has two purposes: (1) to study the usage of large data to design hispeed train service and (2) to create a prototype model for hispeed train service from the master country to get the most benefit from applying for service in Thailand. The Country that has been brought to study for modeling is Japan. The data used in this research will consist of all hi-speed rail service information that is currently available and demographic. These data process by using big data technology and machine learning to help study. After studying we selected 2 suitable models to be used for creating prototype models: multiple linear regression and regression tree. The model predictive performance shows that the multiple linear regression model, which uses the mean square error to test the comparison between the baseline. This model presented in this research is significantly more accurate than baseline with confidence at 95% and from cross-validation in the training dataset. The best predictions have found that the data group consists of demographic data and service information. This model can accurately predict the amount of transportation per day. The findings from this study may directly benefit Thailand’s hi-speed train planners in their effort to develop an optimal train schedule and reduce the future train fare in Thailand.