股指波动性的多重分形谱方法研究  

Volatility Research of Stock Index Based on Multifractal Spectrum Methods

作　　者：马丽[1] 王宏勇[1] 

机构地区：[1]南京财经大学应用数学学院,江苏南京210003

出　　处：《南京财经大学学报》2010年第2期45-49,共5页Journal of Nanjing University of Finance and Economics

基　　金：江苏省高校自然科学基金项目(07KJD110065)

摘　　要：多重分形谱可以分析金融时间序列的微观结构及其特征。本文以上证综合指数的5min高频数据为研究对象,运用多重分形谱分析法考察股指在四种不同情形下的波动状态。通过分析参数的变化与股指波动之间的关系,能有效地预测股指未来的短期变化趋势。此外,用该方法对上证综指日开盘指数的波动性进行研究,也取得了较好的预测结果。实证分析表明股指的未来走向与预测日临近时间段的股指波动状况有较大的关系。Multifractal spectrum can analyze the micro structures and characteristic of finance time series.In this paper,we take 5min high frequency trading data of SSE（Shanghai Stock Exchange） as research objects,and discuss the volatility situations of stock index in four different cases using multifractal spectrum methods.By analyzing the relations between the changes of multifractal spectrum parameters and the fluctuations of stock index,we can predict effectively the changes of stock index in a short time.Moreover,we also get a good result when using this method to the time series of daily stock index.By means of the positive analysis,we find that the stock index near the prediction days have more important influence on the forecast of future stock index.

关 键 词：多分形谱 股指 波动性分析 预测 

分 类 号：F224.9[经济管理—国民经济] F830.9