基于在线理论的股票算法交易策略研究

运用在线理论研究多支股票算法交易策略。在El-Yaniv等人研究基础上,构造了单支股票买入问题的在线策略,证明该策略为最优在线策略;将构造的单支股票交易策略应用到多支股票交易策略问题中,设计了多支股票交易策略算法,并以每支股票收益加权进行投资组合;最后选择上证A股二十支股票从2009年到2012年的交易时间价格数据验证本文所提策略有效性。将20支股票随机抽取10支组成一组,选4组分别进行验证,结果表明本文所给策略对于任意选择的多支股票有较好收益。对交易周期分别选取10个偶数长度进行验证,发现交易周期为18天时平均收益最大,平均收益率为5.2%。

Study on the Stock Algorithmic Trading Strategy Based on Online Theory

The online theory is used to study multi-stock algorithmic trading strategy. On the basis of El-Yaniv’s research, online buying strategy is established and proved to be the optimal online strategy; multi-stock algorithmic trading strategy is designed and the investment portfolio is determined by weighting every stock yield with applying single stock trading strategy into multi-stock trading strategy. Transaction time data of twenty stocks, which are picked out of the A Stock of Shanghai Stock Exchange, is selected to test and verify the validity of the strategy mentioned in this paper. Ten stocks are randomly picked out of these twenty stocks to compose a group, and four groups are selected to be tested respectively, and the result indicates that the strategy proposed in this paper has better yield to any multi-stock. As for transaction cycle, ten even length is selected for test and the result implies the average yield will reach its maximum when the transaction cycle is eighteen and the average yield is 5.2%.