A segmentation based prediction model for expressway toll in regional central cities

Accurate toll prediction is a concern for expressway operating companies. However, due to factors such as time period, week sequence, and weather, the toll on expressways often exhibits extreme imbalance, which also makes accurate prediction of toll very difficult. In response to this issue, this paper proposes a segmented expressway toll prediction model. Using the toll anomaly detection algorithm, the training samples are divided into multiple sub sample sets, and the toll is mapped to the most relevant features. Then, the models are trained separately based on the partitioned dataset. The experiment shows that the regional central city expressway toll model established based on the method presented in this paper has a much better performance than similar traditional models.