Research And Development Of Stock Data Analysis System Based On Time Series Clustering Analysis

Entering the 21st century,with the enhancement of China's economic strength,people have more and more assets in their hands.In order to maximize the income of the assets in their hands,people choose stocks to invest.Stock investment,the first to bear the brunt of its asset allocation and risk control,which is the most concerned issues.Since the birth of stocks,stock price forecasting has become a "deep water area" for people and scholars to overcome difficulties.There are a lot of data in the stock market,among which the trading data and stock quotation are the key,while investors pay more attention to the stock quotation,including opening price,closing price,trading volume and other data.The price of stock is continuously changing.It has the characteristics of periodicity,randomness and tendency.There are some relations and certain rules in it.Investors are concerned about the hidden law of stock data,in order to achieve the maximum return with the minimum risk.And machine learning,data mining as an excellent way to deal with data is exactly what we need.In the field of finance,experts determine investment and risk control by studying the similarity and clustering of financial assets.Through using the financial time series clustering method to analyze the stock data and research progress,it provides investors with strong support in decision-making.This paper makes a brief introduction to the related theoretical knowledge of time series clustering method,and then uses time series clustering analysis to predict the stock data.According to the actual demand of stock forecast,the KDJ index is used to provide mathematical statistics information.Considering the efficiency of the algorithm and the accuracy of stock forecast,the two algorithms of FarthestFirst and EM are used in this paper,and good prediction results are obtained in the experiment.On this basis,the demand analysis,architecture design and implementation of the stock data analysis system based on time series clustering analysis are carried out,and then the function and performance of the system are tested.