Some Studies On Nonlinear Analysis Of Time Series

Precipitation has a dramatic effect on agriculture and industry that it takes a significant role in human's activities. Because of the effects of many nondeterministic factors the precipitation series show great nonstationarity and complexity. This paper aims at using DFA to investigate the correlation properties of the precipitation series and studying the chaos properties of the series based on the character and the estimate methods of chaos so as to provide some basic theory for further study of prodiction on long-range correlation and chaos.In this paper, based on DFA we add some correlated trends to the original precipitation series that a crossover which is independent of diversified nonperiodic trends is obtained and we find a scaling property varying process which may show the chaos property of the series as well. SVD comes into our paper because DFA is effectless to periodic and quasi-periodic trend. And then we study the chaos property of the precipitation seires with DFA. These three parts contact each other with the arithmetic of DFA. Based on the topics above, this paper consists of three parts as below:Chapter1: Application of DFA on the study of long-range correlation of pecipitation seires.Chapter2: Effect of SVD on removing periodic and quasi-periodic trends based on random signals.Chapter3: Analysis of chaos of the precipitation series.