Analysis And Prediction Of Road Traffic Accident Severity Based On Bp Neural Network

Since China’s economic development is very rapid,the level of motorization is becoming higher and higher,and the accidents are becoming more and more frequent,which has greatly affected our life and property safety.Therefore,many scholars continue to study road traffic accidents,explore the impact of various factors on traffic accidents and provide effective methods to prevent traffic accidents by studying internal laws.For the study of traffic accident prediction,in addition to the method of mathematical statistics,the most commonly used method is mathematical modeling.How to choose a suitable method for traffic accidents,which is a complex problem with uncertainty and nonlinearity,has been the study direction of national and international researchers.In this paper,a mathematical model based on BP neural network is used to build a prediction model.Considering the advantages of BP neural network,it can be effectively combined with other methods to build a model with a good applicability.This paper proposes a prediction model for the severity of road traffic accidents based on different algorithms of BP neural network,and verifies that the BP neural network model based on genetic algorithm is more suitable for analyzing the complex relationship between the cause factors of traffic accidents and the severity of accidents.Secondly,a total of 4958 effective road traffic accident cases from 8 provinces from the year 2011 to the year 2019 are selected as the experimental research objects,and the classification of the severity of road traffic accidents is analyzed and discussed;Based on the data,19 factors which affect road traffic accidents are comprehensively considered,and these logical variables are encoded by one-hot method and used as input variables.Based on the given data,the four categories of road traffic accident grades,including minor accident,general accident,major accident and extra serious accidents,are used as output variables,and the correlation between each factor and the severity classification is analyzed by SPSSAU.Then,based on MATLAB programming,4182 road traffic accident data from the year 2011 to the year 2018 are used to train and build a BP neural network prediction model without genetic algorithm and a genetic algorithm optimized BP neural network prediction model.Finally,using the data of 776 road traffic accidents in 2019,the prediction results of the two models are compared.The results prove that the two prediction models can make accurate and effective predictions,and the genetic algorithm’s BP network has more accurate and effective predictions.Using the model to predict the severity of the accident can prevent the accident in advance or reduce the injury after the accident.