摘要
本文研究了多重分形的统计物理方法,以上证综合指数长达四年的一分钟高频数据为研究对象,计算了实际交易数据的多重分形谱及其特征参数,并确定了权重因子的取值范围。结果表明奇异指数和相应的谱函数作为多重分形谱的重要参数,一定程度上反映了股指本身的变化范围和高低价位出现频率的变化,然而谱函数可以预测股市趋势的断言在沪市并不成立。
In this paper, Method of Muhifractal Statistical Physics was used, And multifractal spectrum was computed with 1 min Shanghai Composed Index data within 4 years. The result showed that the important parameter of muhifractal spectrum which was spectral width and its spectral function value in some degree give the domain of stock index and the occurrence rate of high and low price level. But the assertion that spectral function value can predict the trend of stock market was wrong.
出处
《数理统计与管理》
CSSCI
北大核心
2007年第3期535-542,共8页
Journal of Applied Statistics and Management
基金
国家自然科学基金(70371035)支持
关键词
统计物理
高频数据
多重分形谱
奇异指数
statistical physics
high frequency data
multifractal spectrum
singularity exponent
