Exit Traffic Flow Prediction of Highway Based on TPE-XGBoost Model
Due to the rapid growth of the number of family cars and the development of freight and passenger transport busi-ness,the traffic flow of highway traffic is increasing.On the basis of huge data,it is more difficult to predict the traffic flow of highway exit.In order to accurately predict the traffic flow at the exit of highway roads,this paper proposes a extreme gradient boosting(XGBoost)model based on tree-structured parzen estimator(TPE)hyperparameter optimization to predict the traffic flow at the exit of highway roads.The model uses TPE method to seek the optimal combination of hyperparameters of XGBoost algorithm,which can achieve fast optimization of hyperparameters and accurate prediction of traffic flow.Example verification results show that compared with traditional machine learning models such as gradient boosting decision tree(GBDT),random forest and support vector machine(SVM),the proposed model has a higher degree of fitting and lower error on the real data,with R2 up to 0.985 and RMSE up to 0.95.
TPE hyperparameter optimizationXGBoosthighway traffic flow forecastalgorithm of regression
万方数据
维普
