Evaluation And Prediction Of Pavement Performance Evaluation Based On Data Mining

With the development of information technology,China's road network has begun to move toward the "smart road." Road performance evaluation,prediction and decision making are three very important aspects in road maintenance construction and management.When the road is put into use,under the joint action of various factors such as traffic volume and environment,the pavement structure has suffered various degrees of damage,and the pavement performance has been declining year by year.In order to improve the performance of pavement,extend the service life of the road,and provide scientific support for road maintenance,this paper based on data mining technology,fully researches the evaluation,prediction and decision-making of pavement performance,and lays a foundation for the automation and intelligentization of road maintenance in the future.This paper takes the “Hubei Province Smart Highway Big Data Acquisition and Mining Analysis Platform”project as the research background,and applies data mining technology to pavement performance evaluation and prediction decision.The main research contents of this paper are as follows:(1)In view of some data quality problems that may exist in the existing road pavement performance data,the data is preprocessed,including redundant value processing,missing value repair and abnormal detection of pavement use data by iForest algorithm,thereby improving the original data quality,the pre-processed pavement performance data is stored for later pavement performance evaluation,prediction and decision making processes.(2)Existing pavement performance evaluation indicators When evaluating pavements in different regions,the evaluation results obtained sometimes do not match the actual situation.This paper uses the evaluation model based on information entropy and TOPSIS theory to evaluate the existing pavement performance.Comprehensive evaluation,and compare the model with the extension evaluation model and the evaluation results of the current pavement performance specifications,and obtain the feasibility of the evaluation model,and then improve the accuracy of pavement performance evaluation.(3)Aiming at the existing pavement performance prediction,the single prediction model is easy to lose some useful information.This paper uses the nonlinear regression prediction,gray system GM(1,1)prediction and support vector machine entropy weight combination prediction model to use the existing pavement.The performance is predicted and compared with the prediction results of the three single prediction models,and the validity of the entropy weight combination prediction model is obtained.In turn,the accuracy of road surface performance prediction is improved.(4)In view of the fact that the current grey matter element model is applied to the road maintenance decision,the expert weight is basically used,and the expert weight is easily affected by personal subjective factors.Based on the information entropy principle,the weight of the decision indicator is determined,and the road segment maintenance decision-making optimization is optimized for the relevant road segment.