Travelers' Risk Attitude Classification Method Based on Cumulative Prospect Theory and Experimental Results

Attitudes towards risk play an important role in travelers' mode choice. Previous research has pointed out that the individual risk attitude varies with different conditions. The traditional risk attitudes classification of dividing people into three categories, i.e. risk averse, risk neutral and risk seeking, cannot reflect the variation of travelers' risk attitudes. In this paper, we adopt the cumulative prospect theory (CPT) and propose a new risk attitudes classification by dividing people into five categories. Based on our experimental data, we found that the new classification can correctly describe the variation of respondents' risk attitudes when facing gain and loss and all respondents can be divided into different categories representatively and reasonably. We then set a CPT-based mode choice model and calibrated the model based on the experimental data. We found that the influences of the different risk attitudes to parameters were mainly reflected in the value function instead of the weight coefficient. Also, the graphs of the value function associated with different categories of respondents have been plotted to help the data analysis and the new classification has been verified that it can significantly improve the forecasting performance of the CPT-based mode choice model.