Research on SVM-Based Highway Traffic Safety Evaluation Model

Abstract

To enhance the traffic safety level of highway, the paper proposes SVM-based highway traffic safety evaluation model. According to the screening principle and grading standard of evaluation index, and establishes road traffic safety evaluation index system including driver, vehicle, road, traffic management and supervision and determines weight of each index using AHP. Through matter element judgment method, the paper selects the learning sample of SVM evaluation model, constructs SVM-based highway traffic safety evaluation model, solves SVM evaluation model by using SVM multi-classification algorithm of partial binary tree, and optimizes parameters by using GA. Through collecting related data about traffic safety of Luoluan Highway in He’nan and combining the overall evaluation of such Highway using the model, the paper verifies the reliability of model.