沪深300股指期货跨期套利算法研究与系统设计

发布时间：2018-01-04 04:31

本文关键词：沪深300股指期货跨期套利算法研究与系统设计　出处：《东华大学》2013年硕士论文　论文类型：学位论文

【摘要】：股指期货是以股票价格指数为交易标的的期货,投资者运用股指期货,既能够有效的规避股票市场的系统风险,又能够利用它进行套利。目前,股指期货套利功能引起了国内外投资者的广泛关注。因此,本文采用统计套利的思想,对沪深300股指期货的统计套利的程序化交易算法及系统实现进行了研究。本文首先介绍了股指期货跨期套利策略及模型,在此基础上,构建了程序化交易策略的算法,并采用沪深300股指期货当月合约和下月合约的1分钟高频数据进行实证分析,最后用Matlab软件编程实现了套利交易机会的自动判断和收益率计算。在程序化交易策略的算法方面,提出了两种程序化交易策略；之后,在实证方面,首先对配对合约样本内数据进行了协整关系的检验,证明合约间价差存在均值回归,并利用协整检验的回归方程得出的系数作为交易头寸的比例,回归后的残差序列作为套利判断的依据,在正态分布的基础上给出交易的开仓、平仓和止损时点,建立套利交易的策略,计算了交易次数和累计收益率。在Matlab实现方面,考虑期货价格序列方差时变特性,对样本外一个长时段数据采用了移动窗口预测法来预测价差序列的波动,通过Matlab开发的套利系统实现了实时读取数据判断交易,完成程序化套利的过程,并得到样本外数据实证结果,通过对比,移动窗口预测法得到的窗口大小与移动频率决定了套利机会和累计收益率,在经过对这两项参数的优化后取得了良好的套利效果,并有略高出样本内数据的平均收益率
[Abstract]:Stock index futures are futures with stock price index as the trading target. Using stock index futures, investors can effectively avoid the systematic risk of the stock market, but also can use it to carry on arbitrage. The arbitrage function of stock index futures has attracted the attention of investors both at home and abroad. Therefore, this paper adopts the idea of statistical arbitrage. This paper studies the program trading algorithm and system realization of statistical arbitrage of CSI 300 stock index futures. Firstly, this paper introduces the strategy and model of intertemporal arbitrage of stock index futures, and on this basis. The algorithm of programmed trading strategy is constructed, and the high-frequency data of Shanghai and Shenzhen 300 stock index futures are used for empirical analysis of one-minute high frequency data of month contract and next month contract. Finally, the automatic judgment of arbitrage trading opportunity and the calculation of profit rate are realized by using Matlab software, and two programmed trading strategies are put forward in the algorithm of programmed trading strategy. Then, in the empirical aspect, we first test the cointegration relationship of the data in the paired contract sample, and prove that there is a mean regression of the price difference between the contracts. The coefficient obtained by cointegration test is used as the proportion of trading position, and the residual sequence after regression is used as the basis of arbitrage judgment. On the basis of normal distribution, the time points of opening, closing and stopping losses are given. The strategy of arbitrage trading is established, the times of trading and the cumulative rate of return are calculated. In the implementation of Matlab, the time-varying characteristics of the variance of futures price series are considered. For a long time data outside the sample, the moving window prediction method is used to predict the fluctuation of the spread sequence. The arbitrage system developed by Matlab realizes the real-time reading data to judge the transaction. Complete the process of programmed arbitrage and get the empirical results of data outside the sample. By comparison, the window size and frequency of mobile window prediction determine the arbitrage opportunity and cumulative rate of return. After the optimization of these two parameters, good arbitrage effect is obtained, and the average rate of return is slightly higher than that of the data in the sample.
【学位授予单位】：东华大学
【学位级别】：硕士
【学位授予年份】：2013
【分类号】：F224;F724.5;F832.51

【参考文献】