Sliding windows and persistence

Sliding windows and persistence
复制标题

滑动窗口和持久性

DOI:
10.1121/1.4987655
复制
发表时间:
2017
影响因子:
2.4
通讯作者:
Chris Traile
Chris Traile
中科院分区:
物理与天体物理3区
文献类型:
--
作者:
Jose A. Perea;Chris Traile

文献摘要

被引文献

相似文献

在过去的几年里,几何和拓扑思想作为解决信号分析问题的一种手段已经急剧增加。本次演讲的目标是展示如何从动态系统的想法(例如,时间延迟嵌入)与来自拓扑数据分析的工具(例如,持续同源性)允许人们从向量值时间序列数据中提取高度非平凡的特征。作为一个例子,我们描述了一个范例,用于量化(准)周期性的视频数据,并提供了应用程序,包括研究基因调控网络在生物学中,在哺乳动物中的双音化,在人类中的语音病理学。
The use of geometric and topological ideas as a means to tackle problems in signal analysis has seen a sharp increase in the last few years. The goal of this talk is to show how ideas from dynamical systems (e.g., time delay embeddings) with tools from topological data analysis (e.g., persistent homology) allows one to extract highly non-trivial features from vector-valued time series data. As an example, we describe a paradigm for quantifying (quasi)periodicity in video data and provide applications including the study of gene regulatory networks in biology, biphonation in mammals, and speech pathologies in humans.