Futures Data Process System Design And Prediction Analysis

With the continuous development of Internet technology and the popularity of user groups,great changes have taken place in the futures market,based on people's demand for convenience and flexibility of the trading,and detailed requirements for trading strategy,the original manual trading method already can't satisfy the trend of development for the stock market trading.In recently,the Shanghai futures exchange issued an standard open interface for futures trading,which makes the individual access to free exchange interface and function for self-developed programming trading become a reality.Base on this,the dissertation complete a time series data processing system,the main functions of which are interface display,parameter choice personalized,trading strategies established by test using historical data,automated trading,etc.This dissertation try to explore the trend of programming trading.Data mining is more and more widely used on the research of the futures market,and a large amount of data are collected personally by the system above-mentioned.It is significant in how to handle these data and to do some works for forecast.But because of the influence factors of the futures market are complex,so the time series of futures data have the characteristics of not smooth and non-linear,the research base on traditional data mining algorithms have some limitations.So to improve the traditional data mining algorithm and application in futures time series analysis and prediction is very important.Firstly,the dissertation introduce the comprehensive trading platform and the uniform interface provided by the Shanghai futures exchange.On the basis,analyse the system in detail from the aspects of business and function requirement,and design overall structure of system and each function module in detail.Next expound the concrete realization of the system,include the database design for the system,the realization of classes in the system,and the display of a part of the system development work,and etc.And then,introduce the relevant theories of data mining used for futures time series prediction,include the density clustering algorithm and support vector machine regression algorithm.At last,improve the density clustering algorithm base on the futures market having the characteristics that the data is not smooth and have noise,and use the improved density clustering algorithm and support vector machine regression algorithm to forecast time series for futures,and analyses the prediction result.Experiments show that the improved algorithm for prediction on the futures time series can capture the trend of them.