Prediction And Analysis Of The Impact Of Transportation On Urban Carbon Emission Based On PSO-BP Neural Network

In recent years,with the continuous advancement of urbanization,the proportion of carbon emissions from urban transportation in regional carbon dioxide emissions has been increasing year by year.In particular,most of China’s first-tier cities have become the main component of urban carbon emissions.Therefore,the study of the impact of urban transportation on carbon emissions is of great significance to urban low-carbon transportation construction and economic development.On the basis of reading a large number of domestic and foreign relevant literature and summarizing the research status of traffic carbon emission,this thesis obtains specific traffic factor data through a series of data analysis and processing on the basis of urban road network data and taxi track data,and takes Chengdu 6,000 m * 6,000 m grid unit as the research object.A carbon emission neural network prediction model is established to predict and analyze urban carbon emissions and evaluate the results.The specific content is divided into the following aspects:1.Estimation of regional carbon emissions based on light data simulation.The light data were processed and different light factors were extracted.Geographicweighted models with different combination of factors were established to estimate carbon emissions,and the optimal models were compared and evaluated to screen out the optimal model.Finally,the spatial analysis of carbon emissions data estimated by the optimal model was carried out.2.Extraction and analysis of traffic factors.The basic road network data and taxi track data are processed to extract traffic factors and carry out spatial analysis for each traffic factor.Finally,geographical detector is used to analyze the explanatory power of single traffic factor on regional carbon emissions and the change of explanatory power of traffic factor on regional carbon emissions under the interaction of two factors.3.A neural network prediction model of regional carbon emission was constructed based on traffic factors.PSO algorithm to optimize the BP neural network model is established with the standard neural network model,through building the appraisal index of different model performance,degree of fitting and generalization ability three aspects to evaluate the two models,finally,for further verification of model generalization ability,in different spatial scales of regional carbon emissions.In this thesis,the research results show that traffic construction scale and real time traffic status of chengdu has distinct effect on the carbon emissions to the appropriate area,in terms of spatial distribution,the greater the traffic construction scale and real time traffic area,the higher the carbon emissions significantly,and the traffic of Chengdu and carbon distribution center ring radioactive structure characteristics and "high" east west low marked characteristics;The constructed BP neural network model optimized by PSO algorithm is obviously superior to the standard BP neural network model in terms of model performance,accuracy and stability.The prediction accuracy is 95.11% in the space scale of 36km2 grid,and the average prediction accuracy is 92.4%in the larger space scale.The prediction results are accurate and reliable.