Research To Our Country Stock Market Basing Onclustering Analysisand Principal Component Analysis

China's stock market comes into being from scratch and now has been a considerable size.Stock market gradually improved and the intrinsic value of the return is the future direction of the stock market development.Therefore,blue chip stocks higher investment value will grow increasingly sought after by investors.Speculation in the past that high prices a serious deviation from the value of the phenomenon will be gradually corrected.How to choose the different sections in many types of stocks has investment value of the stock has become a serious problem.Faced with large amounts of data information,financial data,fundamental data and the stock price data,how to find useful data is a problem worthy of study data from these big high latitudes of the information.We have entered the era of big data,and data mining and other terms associated with the emergence of more and more come in our field of vision.Cluster analysis as data mining is an important method of application is extremely broad.The basic idea of clustering analysis is based on the similarity between metrics to the data classification,and general similarity matrix structure will determine the quality of the clustering effect,which is easy to intuitively understand.Because if a measure of the relationship between the data measurement more accurate depiction of the results of course,the more it can reflect the relationship between the data We will discover the following data classification still exist many problems to be solved,because when the stock classification is completed,we need to screen out the real investment value of the stock.In response to these problems,this paper selects the Chinese stock market between different sections of 58 stocks.Firstly,we use cluster analysis method to classify the selected sample according to the yields of stocks.Next,the stocks which have homogeneity(whose yield have close contact)is classified as a class.Then,We sort stocks for each individual financial indicators used by the principal component analysis method.So that we will be able to filter out some stocks which have investment potential between different categories.This paper uses the K-means clustering methods and spectral analysis clustering methods,and compares the differences between the two clustering methods.Besides,since the stocks between different categories have different earnings characteristics,which has reduced the risk of investment when choosing stocks from different categories.The so-called do not put your eggs in one basket.This has far-reaching significance for investors...