时间序列分维数提取算法的研究

混沌时间序列的分维数是描述其混沌特性的一个重要参数．这里介绍的改进的分维数提取算法是在传统的Ｇ－Ｐ关联维数分析算法的基础上，结合奇异谱分析技术，构造了基于奇异谱分析的关联维数提取算法，克服了原始Ｇ－Ｐ算法抗噪声干扰能力较差的缺点，同时具有可靠性较高、结果精度较理想、运算速度较快等优点

Study on Estimating the Fractal Dimension of Time Series

Fractal dimension of chaotic time series is an important parameter describing its chaotic property. A modified algorithm is presented for estimating the fractal dimension of time series, which is based on the traditional G P algorithm and combined with Singular Spectrum Analysis (SSA). The modified algorithm, termed as Correlation Dimension Estimating Algorithm Based on Singular Spectrum Analysis, is superior to the traditional one because of its greater robustness to noise, higher stability and ability to yield more accurate results.