Based On The Principal Component Cluster Analysis And The Shenzhen Stock Market Financial Characteristics Of GARCH Model Analysis

Along with our country securities trading becoming more mature,people's awareness is also growing for stock.About all aspects of the stock market has attracted the attention of people.For example,for the course of the stock option and stock price.Clustering analysis as a method of data processing because of its simple and easy to understand the characteristics of widely attention,to help investors in large Numbers,a wide variety of data information,selected for their valuable data,and can choose the most suitable methods for handling.GARCH model is a description of the stock market volatility often used to model.Engle in1982 gives the autoregressive conditional heteroscedasticity(ARCH)model.Up to now,the ARCH model has become one of the most widely used analog sequence model of financial revenue.The ARCH model can depict the distribution of some of the characteristics of financial returns series,but there are also some limitations.On one hand,large,the conditions of the ARCH model variance associated with the start of the variance,easy to make error is bigger;Request parameter is positive,the ARCH model,on the other hand,to ensure that the variance is greater than zero,but in practice,often will encounter parameters being negative.Generalized autoregressive model is put forward accordingly,Bollerslev(GARCH)model.Then,people gradually EGARCH and GARCH-M model is put forward.This article first chapter to clustering analysis in this paper,and introduces the method of principal component analysis.The second chapter introduces the related concepts of volatility model.The third chapter is the empirical analysis,first of all,using GARCH models on the yield of the Shenzhen composite index volatility has carried on the empirical study,used for eviews software,get the EGARCH(1,1)model is better than GARCH(1,1)model and GARCH-M model can better fit the conclusion of the Shenzhen composite index yield fluctuation.Then,again make GARCH(p,q)parameters change,get the EGARCH(1,3)better simulation.In addition,using SPSS software,for 20 randomly selected Shenzhen stock to collect a variety of indicators,according to the principal component analysis method to delete selected first,then using the method of cluster analysis of recent deep 20 stocks are classified and concluded reference for investors.