Prediction of bus travel time is a crucial tool for passengers. We present methods to predict bus travel time over intervals between pairs of adjacent bus stops using city bus probe data. We apply Gradient Boosting Decision Trees to several kinds of features extracted from the probe data. Experimental results illustrate that adding a combination of features improves the accuracy of travel time prediction over the target interval. In particular, the method using a combination of the travel time over the interval previous to the target one and the number of stops the bus makes before reaching the target interval has better performance than the other methods which use all the other combinations of four features used in this study.
|Number of pages||12|
|Journal||International Journal of Intelligent Transportation Systems Research|
|Publication status||Published - Jun 2021|
All Science Journal Classification (ASJC) codes