A Minimum Description Length Technique for Semi-Supervised Time Series Classification.
Nurjahan BegumBing HuThanawin RakthanmanonEamonn J. KeoghPublished in: IRI (best papers) (2013)
Keyphrases
- minimum description length
- semi supervised
- mdl principle
- semi supervised learning
- information theoretic
- unsupervised learning
- semi supervised classification
- multi view
- labeled data
- unlabeled data
- active learning
- dynamic time warping
- nearest neighbor
- label propagation
- pairwise
- pairwise constraints
- metric learning
- supervised learning
- subspace learning
- manifold learning
- labeled examples
- sampling methods
- semi supervised clustering
- information theory
- co training
- maximum likelihood
- image analysis
- constrained clustering
- image processing