Q.L. Zhan, Y. Li

Pages: 91-102

Abstract
Aiming at the problem of poor evaluation effect of existing methods, an evaluation method of urban road traffic economic loss based on XGBoost algorithm is proposed. First of all, a set of comprehensive evaluation indexes of urban road traffic economic losses are constructed and the evaluation index data are collected. Then, the weights of each evaluation index are calculated. Based on XGBoost algorithm, the economic loss evaluation model of urban road traffic is established. Finally, the model is trained by adaptive method to achieve the purpose of accurate evaluation. The experimental results show that the recall rate is more than 97.6%, the calculation error rate of each evaluation index weight is less than 5%, and the evaluation accuracy rate of urban road traffic economic loss is more than 96.8%, which is higher than other comparison methods, and the evaluation effect is good.
Keywords: XGboost algorithm; urban roads; traffic economic losses; loss evaluation; subjective and objective weights; objective function