TOTOPO: Classifying univariate and multivariate time series with Topological Data Analysis
share
This work is devoted to a comprehensive analysis of topological data analysis fortime series classification. Previous works have significant shortcomings, such aslack of large-scale benchmarking or missing state-of-the-art methods. In this work,we propose TOTOPO for extracting topological descriptors from different types ofpersistence diagrams. The results suggest that TOTOPO significantly outperformsexisting baselines in terms of accuracy. TOTOPO is also competitive with thestate-of-the-art, being the best on 20 and 40 that TDA-based approaches arerobust to small perturbations in data and are useful for cases where periodicity andshape help discriminate between classes.
READ FULL TEXT