On measuring confidence levels using multiple views of feature set for useful unlabeled data selection.
Thanh-Binh LeSang-Woon KimPublished in: Neurocomputing (2016)
Keyphrases
- feature set
- multiple views
- unlabeled data
- confidence levels
- multi view
- semi supervised
- labeled data
- semi supervised learning
- co training
- single view
- class labels
- feature selection
- classification accuracy
- viewpoint
- d objects
- supervised learning
- training data
- active learning
- random forest
- training set
- feature space
- text classification
- data points
- labeled and unlabeled data
- feature extraction
- multiple features
- learning algorithm
- text categorization
- selected features
- feature vectors
- statistical significance
- feature subset
- training examples
- kernel function
- support vector machine
- range images
- face recognition
- object recognition
- learning process
- computer vision