Positive-Unlabeled Classification under Class-Prior Shift: A Prior-invariant Approach Based on Density Ratio Estimation.
Shota NakajimaMasashi SugiyamaPublished in: CoRR (2021)
Keyphrases
- class labels
- prior knowledge
- multi class classification
- supervised learning
- databases
- text classification
- density ratio estimation
- data sets
- positive and unlabeled examples
- density estimation
- unsupervised learning
- support vector machine
- feature vectors
- training set
- support vector
- image sequences
- decision trees
- knowledge discovery
- data points
- active learning
- feature space
- feature selection