金融时间序列的多重分形分析毕业论文（本科）.doc

金融时间序列的多重分形分析 MULTIFRACTAL ANALYSIS OF FINANCIAL TIME SERIES 指 导 教 师： 申请学位级别：学 士 论文提交日期：2014年6月12日 摘 要 有效市场假说(EMH)是现代金融市场的基础理论，该理论认为市场的价格反映了市场的全部信息，市场价格的波动之间相互独立而且不可预测，收益率服从随机游走，收益率分布服从正态分布或对数正态分布.但是，现实中的种种限制因素决定着这一传统的金融理论有着很大的局限性，实际的资本市场并不是传统理论所描述的线性系统，而是一个非线性的系统，这也意味着分形理论开始应用在金融市场. 分形理论则认为金融市场具有明显的分形结构和尖峰厚尾的分布特征，金融时间序列在一定的标度范围内有着持续性与反持续性的特征，而且不同幅度的波动能够表现出多重分形特征.分形理论比有效市场理论更能有效揭示金融市场的波动本质，同时也能更有效地揭示出金融市场的基本规律. 本文选取上证综指（上海证券综合指数）和深证成指（深圳证券成分指数）2005年1月5日至2014年5月22日的每日收盘价的股指收益数据位样本，分别采取R/S、DFA、MF-DFA方法对我国股市的分形及多重分形特征进行实证研究与分析.主要验证了两时间序列的分形及多重分形特征；分析比较了两时间序列的市场有效性特征，通过计算并比较的大小，得出了上海证券市场比深证证券市场有效；分析比较了两时间序列的市场风险，通过计算并比较多重分形谱的宽度，得出了上海证券市场存在的风险比深证证券市场的要大. 关键词：分形； 多重分形； 广义Hurst指数；市场有效性； 市场风险 ABSTRACT Efficient Market Hypothesis (EMH) is the basis of modern finance theory, the main idea of EMH is that the financial market prices presents all information of market, fluctuation of market price are not only independent but also unpredictable, the returns follow a random walk hypothesis, and the distributions of the returns is normal or logarithm normal distribution. Yet many abnormal financial visions in reality means that the traditional financial theories have great limitation, it shows that the actual capital market is not a linear system which as the traditional theory described, but a nonlinear system. This also means the appearance and development of fractal theory. The basic view of fractal theory is that the financial market has obvious fractal structure and fat tail characteristics. The financial time series is persistent and anti persistence in a certain scale, different amplitude fluctuations can show multi fractal characteristics. So the fractal theory can reveal the volatility nature more accurately than that of traditional capital market theory, and can effectively reveal basic law of the finance market. This thesis chooses the stock return data on the day closing price between January 5, 2005 to May 22, 2014 of the Shan