【摘要】 针对工程实践中大量分布、因果不清,产生于复杂系统的序列数据,这类序列数据由于生成机制不明或者复杂,无法通过还原论进行分析,同时由于复杂系统的涌现性数据往往波动较大,通过传统的回归等统计学方法往往无法得到有效分析,本文讨论了CUP算法的在序列拟合上的应用,提出了CUP分析的一般过程和方法,并就因果发现、序列模式识别等方面结合仿真实验进行了讨论,并就一般特性的差异比较了常见的因果发现和模式识别方法,对工程实践中的复杂系统背景的序列数据分析有指导作用。更多还原

【Abstract】 There is a large number of data sequence with unclear distribution and causality in engineering practice,it is generated from the generation mechanism of unknown and complex system. Such data sequence cannot be analyzed by reductionism. At the same time the data is fluctuating frequently.Through traditional regression and other statistical methods are hard to obtain effective analysis. This paper discussed the application of CUP algorithm in sequence fitting and proposed the general process and method of CUP analysis of causal discovery and sequence pattern recognition. This common methods of causality discovery and sequence pattern recognition combining with experiment were compared for discussing the differences in general characteristics,which provided guidance for the analysis of sequence data in complex system backgrounds in engineering practice.更多还原