讲座题目：Graph-let BasedTime Series Analysis

主讲人：Huijie Yang (杨会杰)

开始时间：9月6日9:30-10:30

讲座地址：闵行校区物理楼226报告厅

报告人简介：

上海理工大学系统科学系，教授，博士生导师。长期致力于复杂系统状态描述与危机早期预警信号研究，涉及短时间序列分析、基于复杂网络的时间序列分析和复杂网络的谱结构等。在IEEE Trans Neu Sys Reha Eng, PRE, JTB等刊物发表SCI论文80余篇。承担和参与科技部973，自然科学基金重点和面上项目多项。中国系统科学学会理事。上海市高水平大学建设项目，上海理工大学《复杂系统分析与早期危机预警战略团队》负责人。

报告内容简介:

A complex system is consisted of many elements which are networked by thecomplicated relationships. Records of the dynamical process of the system forma mono/multi-variate time series. There exist non-trvial patterns in theseries. And the patterns are inter-dependent on each other, rather thanindependent. The patterns and their relationships are determined by theunderlying dynamical mechanics, and can take subsequently as representations ofthe system’s state and its evolution. In the present talk, I will introduce ourrecent works on the distribution behaviors and interdependent networks ofpatterns in stochastic and deterministic processes. What is more, I will alsointroduce our work on structure of cross-correlations between time series,which is merged completely by the staistical procedure of average in classicalstatistical methods.

复杂系统诸多元素之间通过其间复杂的关系构成网络结构。复杂系统的动力学过程的输出记录，构成一个（单）多变量时间序列。时间序列中存在不同的花样结构，花样结构之间通过其依赖关系形成花样网络。这些花样结构及其网络隐含着复杂系统动力学特征，因而可以作为复杂系统状态及其演化的描述。本报告将介绍本课题组在这方面的一些工作，包括随机和确定性过程状态分布特征及其演化，序列互关联的结构等。