Change-point Detection And Analysis Of Brent Crude Oil Based On WBS Model

Change-point detection is a basic and hot issue in many fields,such as image analysis and biomedical in recent years,and it has a wide range of applications.In this paper,on the basis of the standard BS algorithm,we introduce a novel algorithm using the idea of random localization mechanism named Wild Binary Segmentation(WBS),for consistent estimation of the number and locations of multiple change-points in data by calculating the CUSUM statistics of each extraction interval to determine the significance of strain points.This paper starts from the basic principles of WBS and data simulation,the purpose is to improve the learning efficiency of algorithm and time series model,and explore the effectiveness and feasibility of the whole process.And try to better improve the accuracy of fitting by changing parameter choice,thus we can reduce the computation time and detect the sequence more accurately.Moreover,we propose two stopping criteria for WBS,i.e.threshold test and information criterion test.This paper focuses on the case analysis of Brent crude oil.Through a series of theoretical analysis and simulation results,it shows that the WBS algorithm has an excellent detection performance in the research fields.It not only greatly enhances the learning efficiency and computational accuracy,but also uses reasonable knowledge to analyze and predict the research sequence,and shows a good fitting degree.At the same time,the discussion of crude oil price is of great practical and theoretical significance to China's economy development.