Time series cluster kernels to exploit informative missingness and incomplete label information.
Karl Øyvind MikalsenCristina Soguero-RuízFilippo Maria BianchiArthur RevhaugRobert JenssenPublished in: CoRR (2019)
Keyphrases
- label information
- missing data
- missing values
- semi supervised
- dimensionality reduction
- distance metric
- unlabeled data
- labeled data
- training data
- data points
- class labels
- multiple types
- transfer learning
- feature space
- active learning
- geometric structure
- kernel function
- kernel methods
- real world
- support vector
- machine learning
- domain specific
- co occurrence
- decision trees
- data sets