Asymmetric learning vector quantization for efficient nearest neighbor classification in dynamic time warping spaces. (April 2018)

• Record Type: Journal Article
• Authors: Jain, Brijnesh J.; Schultz, David
• Is Part Of: Pattern recognition. Volume 76(2018:Apr.)
• Page Start: 349
• Page End: 366
• Publication Date: 2018-04
• Subjects: Learning vector quantization -- Time series -- Dynamic time warping; Pattern perception -- Periodicals; Perception des structures -- Périodiques; Patroonherkenning; 006.4
• DOI: 10.1016/j.patcog.2017.10.029
• Languages: English
• ISSNs: 0031-3203

Abstract

Highlights: Asymmetric LVQ scheme for time series. Decision boundary defined by two prototypes is piecewise quadratic. Margin growth principle for LVQ methods in arbitrary distance spaces. Empirical comparison of prototype generation methods for NN classification. Asymmetric generalized LVQ best trades speed against accuracy. Abstract: The nearest neighbor method together with the dynamic time warping (DTW) distance is one of the most popular approaches in time series classification. This method suffers from high storage and computation requirements for large training sets. As a solution to both drawbacks, this article extends learning vector quantization (LVQ) from Euclidean spaces to DTW spaces. The proposed generic LVQ scheme uses asymmetric weighted averaging as update rule. We theoretically justify the asymmetric LVQ scheme via subgradient techniques and by the margin-growth principle. In addition, we show that the decision boundary of two prototypes from different classes is piecewise quadratic. Empirical results exhibited superior performance of asymmetric generalized LVQ (GLVQ) over other state-of-the-art prototype generation methods for nearest neighbor classification.