基于多重分形的量化投资策略研究.pdfVIP

摘 要 上世纪九十年代分形市场假说的提出对有效市场及其正态假设提出了有力 的挑战。分形市场具有明显的非线性特征，如自相似性、标度不变性、长程相关 性等，而多重分形又是单重分形的推广和补充。以往的研究对刻画金融市场的分 形特征以及利用单重分形进行量化交易都做出了一些尝试。本文基于上证综指和 深证成指的日收盘价数据，使用K-S 检验和JB 统计量论证了两个市场收益率序 列的非正态性，随后从分形的角度来描绘中国资本市场的内在动力学特征，使用 多重分形去趋势波动分析（MF-DFA）方法验证了上海证券市场和深圳证券市场的 多重分形特性，并计算出多重分形谱。 多重分形谱的存在及其左右不对称的性质既说明了金融时间序列的多重分 形特性，体现了资本市场波动的内在规律，又从细节上刻画出波动的大小及其发 生的频率。利用金融时间序列的这一重要性质，本文构建了基于动量效应的简单 趋势策略、基于单重分形的Hurst 指数策略以及加入多重分形谱参数的改进的择 时策略并进行对比分析，发现多重分形谱参数的加入使量化交易策略在金融市场 剧烈波动的时间段内取得了较好的效果，能够抵御市场风险并取得超额收益。最 后，使用统计检验中的自举法，本文发现了量化交易策略在真实预测力上仍然存 在改进的空间。 关键词： 多重分形谱、量化择时、趋势策略、自举法 I Abstract The introduction of the fractal market hypothesis in the 1990s posed a powerful challenge to efficient markets and their normal assumptions. The fractal market has obvious non-linear characteristics, such as self-similarity, scale invariance, long-range correlation, etc., and multifractal is the promotion and supplement of single fractal. Previous studies have made some attempts to characterize the fractal characteristics of the financial market and use single fractals to quantify transactions. Based on the daily closing price data of the Shanghai Composite Index and Shenzhen Component Index, this paper uses the KS test and JB statistics to demonstrate the non-normality of the two market return series, and then describes the internal dynamics of the Chinese capital market from the perspective of fractal According to the characteristics of multi-fractal detrending flutuation analysis (MF-DFA), the multi-fractal characteristics of Shanghai and Shenzhen stock markets were verified, and the multi-fractal spectrum was calculated. The existence of multi-fractal spectrum and its asymmetrical nature both explain the multi-fractal characteristics of financial t