基于支持向量机的股市风格轮换策略研究

作为一种增加潜在收益的工具,风格轮换策略已经成为广为接受的金融投资策略.通过对几种常见的风格轮换模型进行比较,采取支持向量机分别建立规模溢价和价值溢价的预测模型,根据预测结果形成规模风格和价值/成长风格轮换策略,并检验这种积极风格轮换策略在中国股市的有效性.研究结果表明,中国股市存在显著的价值溢价,规模风格轮换策略业绩显著优于消极持有小盘股策略和市场收益,适合采取规模风格轮换策略进行短期投资;消极持有价值股的投资策略业绩显著优于规模风格轮换的投资策略,这暗示在长期投资上更适合采取买入并持有价值股的消极风格投资策略.

An Empirical Study on Style Rotation Strategy in Stock Market Based on Support Vector Machine

As a tool to improve potential investment performance, style rotation strategy has become a widely accepted financial investment strategy. How to model the style rotation strategy has been an importance issue concerned by the academic and practi-cal professionals. Compared with other common style rotation model, we use the latest statistical learning methods--Support Vector Machine to establish the size premium and value premium forecast model. Based on the result from the forecast model, this paper makes an empirical analysis to test the feasibility of the style rotation strategy in Chinese stock market during the period January 2006 to March 2009. The result shows that there is a significant value premium in stock market. The size style rotation strategy can be superior to small cap holding strategy and the stock market. It is suitable to take size style rotation as short-term investment strategy. Our empirical analysis also shows that the fixed value-oriented style strategy can perform better than the val-ue style rotation strategy and small cap holding strategy. This suggests that weshould take the fixed value-oriented style strategy as long-term investment strategy.