Sinkhorn Divergence of Topological Signature Estimates for Time Series Classification

Abstract

Distinguishing between classes of time series sampled from dynamic systems is a common challenge in systems and control engineering, for example in the context of health monitoring, fault detection, and quality control. The challenge is increased when no underlying model of a system is known, measurement noise is present, and long signals need to be interpreted. In this paper we address these issues with a new non parametric classifier based on topological signatures. Our model learns classes as weighted kernel density estimates (KDEs) over persistent homology diagrams and predicts new trajectory labels using Sinkhorn divergences on the space of diagram KDEs to quantify proximity. We show that this approach accurately discriminates between states of chaotic systems that are close in parameter space, and its performance is robust to noise.

arXiv comment9 pages, 4 figures, 2018 17th International Conference on Machine Learning and Applications

References in corpus (10)

• Recurrence Plots for the Analysis of Complex Systems
• Near-linear time approximation algorithms for optimal transport via Sinkhorn iteration
• Smooth and Sparse Optimal Transport
• Large Scale computation of Means and Clusters for Persistence Diagrams using Optimal Transport
• Regularized Optimal Transport and the Rot Mover's Distance
• The density of expected persistence diagrams and its kernel based estimation
• Computational Topology Techniques for Characterizing Time-Series Data
• Geometry Helps to Compare Persistence Diagrams
• Sketches of a platypus: persistent homology and its algebraic foundations
• Sinkhorn Divergence of Topological Signature Estimates for Time Series Classification

Cited by in corpus (1)

• Sinkhorn Divergence of Topological Signature Estimates for Time Series Classification