Construction and Application Research of Change Point Detection Model using Importance Sampling based Markov Chain Monte Carlo

The time series is a basically preserved type of the data as well as it is examined in various manner in real-time applications. Through the probable estimation, the change point detection model is the most important illative targets for researching the analysis of a time series. However, conventional approaches for the change point detection continuously affected through various drawbacks, comprises of an incapability to maintain challenge data patterns, compassion to noise as well as complexities in balancing computational efficiency through accuracy. Hence, this research proposes the Importance Sampling based Markov Chain Monte Carlo (IS-MCMC) method for an effective change point detection in the time series. The IS supports to consider the MCMC sampling process on areas of the parameter space which are most significant for the change point detection model. The IS enables the MCMC to focus on the greatest likelihood areas where change points are most likely to arise. The experimental results demonstrates that the proposed IS-MCMC approach attains the better accuracy of 0.934 and Mean Squared Error (MSE) of 0.482 respectively as compared to the existing approaches like Hamiltonian Monte Carlo (HMC) and Slice Sampling (SS) approaches.