Asymmetric Learning Vector Quantization for Efficient Nearest Neighbor\n Classification in Dynamic Time Warping Spaces

2017/03/24 by Brijnesh J. Jain, David Schultz, Jain, Brijnesh +1
Chemistry · Computer Science · Engineering · #Advanced Chemical Sensor Technologies #Anomaly Detection Techniques and Applications #FOS: Computer and information sciences #Machine Learning (cs.LG) #Machine Learning (stat.ML) #Neural Networks and Applications #Spectroscopy and Chemometric Analyses #Time Series Analysis and Forecasting

paper · pdf · doi:10.48550/arxiv.1703.08403

openalex publication_date 2017/03/24 · openalex created_date 2022/10/01 · openalex updated_date 2026/07/28

Abstract

The nearest neighbor method together with the dynamic time warping (DTW)\ndistance is one of the most popular approaches in time series classification.\nThis method suffers from high storage and computation requirements for large\ntraining sets. As a solution to both drawbacks, this article extends learning\nvector quantization (LVQ) from Euclidean spaces to DTW spaces. The proposed LVQ\nscheme uses asymmetric weighted averaging as update rule. Empirical results\nexhibited superior performance of asymmetric generalized LVQ (GLVQ) over other\nstate-of-the-art prototype generation methods for nearest neighbor\nclassification.\n

Citations

• Dynamic programming algorithm optimization for spoken word recognition
• Nonsmooth Analysis and Subgradient Methods for Averaging in Dynamic Time\n Warping Spaces

Related

• Warped-Linear Models for Time Series Classification
• Similarity Learning for Time Series Classification
• An Analytical Approach to Improving Time Warping on Multidimensional Time Series
• Nonsmooth Analysis and Subgradient Methods for Averaging in Dynamic Time\n Warping Spaces
• metricDTW: local distance metric learning in Dynamic Time Warping