A Wasserstein Subsequence Kernel for Time Series.
Christian BockMatteo TogninalliM. Elisabetta GhisuThomas GumbschBastian RieckKarsten M. BorgwardtPublished in: ICDM (2019)
Keyphrases
- dynamic time warping
- kernel function
- support vector
- distance measure
- non stationary
- pattern matching
- kernel methods
- similarity search
- euclidean distance
- pointwise
- longest common subsequence
- kernel regression
- mutual subspace method
- abnormal patterns
- multivariate time series
- kernel learning
- component analysis
- gaussian processes
- autoregressive
- feature space
- reproducing kernel hilbert space
- positive definite
- subsequence matching
- neural network