Analysis on predict model of railway passenger travel factors judgment with soft-computing methods

Purpose: With the development of the transportation, more traveling factors acting on the railway passengers change greatly with the passengers’ choice. With the help of the modern information computing technology, the factors were integrated to realize quantitative analyze according to the travel purpose and travel cost. Design/methodology/approach: The detailed comparative study was implemented with comparing the two soft-computing methods: genetic algorithm, BP neural network. The two methods with different idea were also studied in this model to discuss the key parameter setting and its applicable range. Findings: During the study, the data about the railway passengers is difficult to analyzed detailed because of the inaccurate information. There are still many factors to affect the choice of passengers. Research limitations/implications: The model-designing thought and its computing procession were also certificated with programming and data illustration according to thorough analysis. The comparative analysis was also proved effective and applicable to predict the railway passengers' travel choice through the empirical study with soft-computing supporting. Practical implications: The techniques of predicting and parameters' choice were conducted with algorithm-operation supporting. Originality/value: The detail form comparative study in this paper could be provided for researchers and managers and be applied in the practice according the actual demand.