Channel masking for multivariate time series shapelets
This work solves a domain-specific problem for time series analysis, offering an incremental improvement over existing methods.
The paper tackles the problem of multivariate time series classification by addressing overfitting from noisy channels, proposing a shapelet learning scheme with channel masks that yields improved results.
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.