arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 1806.00381stat.MLcs.LGmath.PRmath.STstat.TH
收藏

拓扑数据分析中的持久化路径和签名特征

Persistence paths and signature features in topological data analysis

University of Oxford(牛津大学);Institute for Advanced Study(高等研究院)

Ilya Chevyrev, Vidit Nanda, Harald Oberhauser

摘要:

我们引入了一个新的特征图的条形码,出现在持久的同源性计算。其主要思想是首先将每个条形码实现为方便的向量空间中的路径,然后计算其路径签名,该路径签名在该向量空间的张量代数中取值。这两个操作的组合-条形码到路径,路径到张量序列-产生具有统计学习的几个期望属性的特征图,例如普遍性和特征性,并在常见的分类基准上实现最先进的结果。

英文摘要:

We introduce a new feature map for barcodes that arise in persistent homology computation. The main idea is to first realize each barcode as a path in a convenient vector space, and to then compute its path signature which takes values in the tensor algebra of that vector space. The composition of these two operations - barcode to path, path to tensor series - results in a feature map that has several desirable properties for statistical learning, such as universality and characteristicness, and achieves state-of-the-art results on common classification benchmarks.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑