Abstract: Reversal online portfolio strategy can make full use of the reversal effect in the stock market, and the backtest results on most data sets show that the reversal online portfolio strategy can obtain higher cumulative returns. Based on the reversal effect of stock price in the financial market, an expert opinion pool is constructed and a competitive online portfolio by aggregating expert opinions is established. First, the reversal effect of stock prices in windows is used to construct investment strategies representing the expert opinion, and expert opinion pools are obtained based on different window lengths. Then, a weak aggregating algorithm is used to assign trust weight to each expert, and an online portfolio strategy is constructed by aggregating expert opinions. The competitive performance analysis proves that the strategy constructed can follow the offline optimal expert opinion. The numerical results show that the performance of the strategy is better than that of online strategies related.

Key words: online portfolio, weak aggregating algorithm, reversal effect, competitive strategy

摘要： 反转型在线投资组合策略通过有效利用股票市场中的反转效应，在多个数据集上的回测中展现出显著的累积收益优势。基于金融市场股票价格的反转效应构建专家意见池，并以此建立集成专家意见的竞争性在线投资组合策略。首先，利用股票价格在窗口内的反转效应构建代表专家意见的投资策略，并基于不同长度的窗口得到专家意见池；其次，运用弱集成算法为每个专家赋予相应的信任权重，集成专家意见形成在线投资组合策略，竞争性能分析表明，该策略能够有效追随最优专家意见；最后，数值分析结果显示，该策略在实现收益等方面的性能均优于现有相关在线策略。

关键词: 在线投资组合, 弱集成算法, 反转效应, 竞争性策略