摘要
本文提出了分析非线性时间序列动力学相似性的模糊相似性指数方法,通过计算不同时间序列或者同一时间序列的不同区段的动力学相似性,可以识别不同动力学系统的内在属性或者同一系统的动力学状态的改变．该方法在计算关联和时,我们用Guassian函数代替Heavyside阶跃函数,克服了Heavyside阶跃函数的刚性边界问题．通过对Logistic映射数值序列和脑电信号实例检验表明,该方法计算相似性指数更加准确和平稳。
In this paper, we proposed a fuzzy similarity index method to analyze the similarity of nonlinear time series. Using this method, different types of systems or different states of a single system can then be distinguished by calculating similarity between different time series or different segments of same time series. Gaussian function is instead of the Heavyside function within correlation sum at calculating a similarity index, so the rigid boundary of the Heavyside function is eliminated. Simulation and real EEG data confirmed that this method was more accurate and robust.
出处
《信号处理》
CSCD
北大核心
2006年第2期168-171,共4页
Journal of Signal Processing
基金
国家自然科学基金(批准号:60274023)资助的课题
