Application Of Threshold Model To The Study Of Anti-cheating Of Vehicle Exhaust Detection

The number of cars in China has increased year by year,since the reform and opening up,the emission of automobile exhaust has seriously polluted the environment.As an answer measure of reducing the harm caused by automobile exhaust,the relevant management departments have increased the supervision of automobile exhaust emissions,however,the car owner have their own counterpart measure of cheating in the detection of exhaust emission.The endless stream of cheating methods makes it difficult to carry out automobile exhaust gas detection work,which greatly increase the management difficulty of the relevant management departments.At the same time,with the increasing number of cars,it not only requires a lot of manpower and material resources,but also consumes a lot of money while relying on traditional methods for anti-cheating inspection.Therefore,statistical methods are applied in this paper to reduce cheating in automobile exhaust gas detection systems.This paper,basing on the existing classification methods and combining the characteristics of automobile exhaust detection data,draws on the threshold model proposed by Bruce E.Hansen in 2004,and uses the threshold model to analyze the 2018 automobile exhaust detection data in an innovative way,and obtains the expected detection effect.The main research contents of this article include:(1)A brief review of the threshold model,which is used for the empirical analysis of 2018 automobile exhaust data,and the estimation of its model parameters.(2)Studying the parameter estimation of the threshold model from the perspective of Bayesian,discussing the selection of the prior distribution of model parameters,and giving the MCMC algorithm for calculating the Bayesian estimation of model parameters,and using it in Empirical analysis of automobile exhaust gas detection data.(3)The AUC of the threshold model obtained by the frequency method is0.7295162,and it’s 0.7295162 by Bayesian method.This shows that the Bayesian threshold model has higher accuracy for the data analysis.