Confidence path regularization for handling label uncertainty in semi-supervised learning: use case in bipolar disorder monitoring.
Kamil KmitaGabriella CasalinoGiovanna CastellanoOlgierd HryniewiczKatarzyna Kaczmarek-MajerPublished in: FUZZ-IEEE (2022)
Keyphrases
- semi supervised learning
- regularization framework
- manifold regularization
- label propagation
- semi supervised
- multi label classification
- unlabeled data
- labeled data
- multi view learning
- high confidence
- supervised learning
- unsupervised learning
- text categorization
- semi supervised classification
- class labels
- multi instance learning
- learning problems
- machine learning
- co training
- labeled examples
- partially labeled
- shortest path
- active learning
- positive and negative
- multi label
- training data
- learning models
- unlabeled instances
- domain adaptation
- graph laplacian
- graph based semi supervised learning
- semi supervised learning algorithms
- data sets
- semi supervised clustering
- graph construction
- label information
- labeled and unlabeled data
- multi instance
- transfer learning
- text classification
- pairwise
- object recognition
- metric learning
- decision trees
- clustering algorithm