Traffic Data-driven Road Congestion Status Determination And Prediction Analysis

The rapid development of urbanization has caused a large number of people to gather in cities.The traffic demand generated by people's travel exceeds the traffic supply capacity of urban roads,causing overloading of transportation infrastructUre.Frequent traffic jams,traffic accidents,traffic pollution,traffic noise,parking,and insuflBcient parking spaces have a serious impact on the quality of life of residents.Therefore,how to make real-time judgment on the traffic state of the road and forecast the road traffic state has become the main problem of research.Based on this situation,this paper uses the designed KFCM-PSO-SVM algorithm model to accurately divide and judge the road congestion state,and uses the improved support vector machine algorithm to predict the congestion determination parameters.Under the premise of ensuring the feasibility of establishing a model,the results of discrimination and prediction are improved and improved.The research results mainly include the following aspects:First,process the raw data collected by the University of Minnesota Traffic Data Research Laboratory.Since the original data is stored in the form of a.traffic file,it needs to be extracted into a.csv file by the python program,and the data is subjected to noise reduction and the like.Then analyze the data characteristics through SPSS software.Finally,introduce the concept and type of traffic congestion,according to the selection principle of the traffic flow parameters,the traffic parameters(vehicle flow and occupancy rate)and the indicators for evaluating the congestion level are determined.Secondly,Aiming at the shortcomings of traditional clustering algorithm in road traffic status determination,the algorithm is improved.The K-means clustering algorithm is proved to be very dependent on the initial clustering center.Therefore,the interval statistical algorithm is used to determine the initial clustering center k of the optimal sample data.Considering that noise and isolated points have certain influence on the robustness of the algorithm,in order to solve this problem,the kernel function is introduced into the fuzzy C-means algorithm,and the kernel function is Gaussian radial basis function,its purpose is to map the original sample data into the space of the feature.The raw data is mapped to high-dimensional space.The problem of non-linearity is transformed into a linear one,which simplifies the problem essentially.Although the dimension of the function is very high in the feature space,when the inner product of the introduced Gaussian radial basis function is smaller than the inner product in the feature space,the calculation speed of the model algorithm can be improved.The core of the KFCM-PSO-SVM algorithm model designed in this paper is to use the improved support vector machine algorithm to strengthen the clustering of the improved fuzzy C-means clustering algorithm,so as to make the accuracy of traffic state determination more accurate;then compared with the results of K-means-PSO-SVM algorithm,fuzzy C-means clustering algorithm and improved fuzzy C-means clustering algorithm,the new algorithm KFCM-PSO-SVM has higher accuracy;then particle swarm optimization(PSO)is used to optimize support vector machine(SVM)for regression prediction of road congestion determination parameters,and the road traffic state is divided.Finally,the model data is verified using the processed data.The results show that the prediction accuracy of the improved new algorithm KFCM-PSO-SVM is 2.4306%higher than that of the K-means-PSO-SVM algorithm,and the accuracy of the improved algorithm is higher.