Detection Methods Based On Semi-parametric Change Point And Their Applications In The Stock Market

The semi-parametric change point detection method usually analyze change points by using empirical likelihood method, and it is widely applied in probability and mathematical statistics. Semi-parameters change points detection methods not only make up for the shortcomings that parameter models need to establish stronger basic assumptions for parameter models, but they also make up for the deficiencies about that non-parametric models cannot make full use of the known information, so this method can be applied in the parameters and non-parametric models, and then it has better applicability.Firstly, the introduction introduces the development status and research tendency,moreover, it gives out the main research theory of the half parameter change point detection method, then the paper discussed the future research direction of change point problem. Secondly, it considers the changing point problems including the popular change point and asymptotic point through empirical likelihood method, and finish the work of comparison the classification according to change point detection methods, objects and the detection numbers, then it proceed to compare the semi-parameter change point detection method and non-parameter detection method by making numerical simulations.Also it discusses the theory of semi-parameters change point detection method, according to the given three hypotheses and under the the three hypotheses, this paper give the concrete proof of the theorems in this paper through the Taylor expansion. Then it studistudies the change point problems contained the same parameters weighting function and the different weight function respectively, through the the numerical simulation it proves the the superiority of parameter change point detection method. Finally, through the study and analysis of the the half parameter change point detection problem, the specific examples are given by applying these to the stock market, according to the analysis of the historical trend of s&p 500 index stock data, furtherly, it gains the conclusion that tested changing point is corresponding to the practical problems corresponding to the practical problems which proves that the parameter change point model is effective.To sum up, the semi-parametric change point detection can analyze the variation of of the stock market. the paper expounds several kinds of change point detection method,at the same time, it proves the related theorems to changing points, according to the given changing point model, it analyses the stock market data, lastly, it verified the validity and superiority of this method.