A Technology Of Change Point Estimation And Its Application Based On Bayesian

Despite the large number of results and their applications in different fields that have been produced through continuous research on change point problems,change point problems remain a hot topic.The article focused on the change-point problem in the Poisson distribution,using the Bayesian method as a basis.Firstly,based on the Poisson distribution,a Poisson distribution single change-point model was established.Under three types of noninformative prior distributions,SPD,LPD,and JPD,the posterior distribution of each parameter in the model was derived respectively,studying the parameter estimation problem of the Poisson distribution scale parameter single change-point model.R language was used to simulate and compare the estimation accuracy of each parameter under different scenarios of the Poisson distribution single change-point problem.The experimental results showed that there were differences in the estimation accuracy of each parameter in the model under three different non-informative priors in different scenarios.Finally,the method proposed in this article was applied to empirical analysis.Afterwards,this article extended the Poisson distribution single change point model to the Poisson distribution multiple change point model,and proposed to use hierarchical prior methods to study and analyze it.In this way,the parameter estimation problem of Poisson distribution scale parameter multiple change point model was studied,and the Bayesian method was used for research and analysis.Through simulation and empirical analysis,the proposed method was compared with the methods previously proposed by others.Both results showed that the estimation accuracy of each parameter in the Poisson distribution multiple change point problem was higher than that under the methods proposed by previous researchers,indicating that the proposed method for estimating the Poisson distribution multiple change-point problem is feasible and has higher accuracy.