Analysis And Research On Urban Traffic Conditions Based On Meteorological Big Data

The rapid development of China's social economy requires a sound urban transportation system.How to make real-time and reasonable analysis of urban traffic conditions is an important link in improving the urban traffic system.According to the research,urban traffic road index is an important indicator reflecting urban traffic conditions,and it has a significant correlation with the urban meteorological factors.Based on this,based on the big meteorological data of cities,combining principal component analysis(PCA),BP neural network,extreme learning machine(ELM)and long-short time memory network(LSTM)and other theoretical models,a regression model of urban road traffic index is established,in order to provide reference for further improvement of urban road traffic system.First,this paper introduces the research background and significance of urban traffic condition analysis,and expounds the current research status of domestic and foreign scholars on urban traffic condition analysis based on meteorological data.Secondly,this paper introduces the source of data,expounds the data preprocessing method,and uses correlation analysis and PCA method to analyze the influence of meteorological factors on road traffic index.Thirdly,this paper expounds two regression prediction models of road traffic index: PCA-BP and PCA-ELM.The simulation results show that both models can achieve better prediction results,and PCA-BP regression model has better prediction results.However,for large road traffic indexes,the generalization ability of the two network models is poor.Fourth,based on the autocorrelation of road traffic index,this paper proposes improved PCA-BP and PCA-ELM prediction models,namely,I-PCA-BP and I-PCA-ELM.Through the model prediction results,it is found that the generalizationability of I-PCA-BP and I-PCA-ELM models is significantly improved compared with PCA-BP and PCA-ELM,and the predicted value of road traffic index is basically consistent with the real value.Finally,in order to make full use of the autocorrelation of road traffic index,a depth learning network model based on PCA-LSTM is established.The prediction results show that compared with I-PCA-BP and I-PCA-ELM prediction models,the model significantly improves the generalization ability of higher traffic index prediction and shows better road traffic index fitting effect.