Construction of Personalized Bus Travel Time Prediction Intervals Based on Hierarchical Clustering and the Bootstrap Method

Providing accurate bus travel time information is very important to help passengers plan their trips and reduce waiting times. Due to the uncertainty of the bus travel time, the traditional prediction value of the travel time point cannot accurately describe the reliability of the prediction result, which is not conducive to passengers waiting for the bus according to the prediction result. At the same time, due to the large differences in the individual driving styles of the bus drivers, the travel time data fluctuate greatly, and the accuracy and reliability of the point prediction results are further reduced. To address this issue, this study develops a personalized bus travel time prediction intervals model for different drivers based on the bootstrap method. Personalized travel time prediction intervals were constructed for drivers with different driving styles. To further improve the quality of travel time prediction intervals, this study optimizes training data sets considering driving style factors. Then, this paper integrates hierarchical clustering, an artificial neural network, and the bootstrap method to construct another prediction intervals model for bus travel time based on driver driving style clustering and the bootstrap method. The real−world driving data sets of the No. 239 bus in Shenyang, China, were used for experimental verification. The results showed that the two models constructed in this paper can effectively quantify the uncertainty of the point prediction results, the PICP of each interval exceeding the confidence level set (80%). It was also found that the quality of the prediction intervals constructed by clustering the driving style data is better (MPIW values decreased by 23.33%, 54.24%, and 28.61 respectively, and the corresponding NMPIW values also decreased by 18.93%, 10.39%, and 14.19%, respectively), which can provide passengers with more reasonable suggestions for waiting time.