Channel masking for multivariate time series shapelets

TL;DR

This paper introduces a novel shapelet learning method for multivariate time series that employs channel masks to reduce noise impact and improve classification accuracy.

Contribution

It proposes a new shapelet learning scheme with channel masks to handle noisy channels in multivariate time series classification.

Findings

Channel masks effectively reduce overfitting caused by noisy channels.

The method improves classification accuracy on multivariate datasets.

Channel masking acts as an implicit regularizer for shapelet learning.

Abstract

Time series shapelets are discriminative sub-sequences and their similarity to time series can be used for time series classification. Initial shapelet extraction algorithms searched shapelets by complete enumeration of all possible data sub-sequences. Research on shapelets for univariate time series proposed a mechanism called shapelet learning which parameterizes the shapelets and learns them jointly with a prediction model in an optimization procedure. Trivial extension of this method to multivariate time series does not yield very good results due to the presence of noisy channels which lead to overfitting. In this paper we propose a shapelet learning scheme for multivariate time series in which we introduce channel masks to discount noisy channels and serve as an implicit regularization.