A Stopping Criterion for Transductive Active Learning.
Daniel KottkeChristoph SandrockGeorg KremplBernhard SickPublished in: ECML/PKDD (4) (2022)
Keyphrases
- stopping criterion
- active learning
- semi supervised
- transductive inference
- unlabeled data
- transductive learning
- semi supervised learning
- labeled data
- convergence rate
- cost function
- machine learning
- training set
- region merging
- random sampling
- selective sampling
- training examples
- merging algorithm
- active learning strategies
- supervised learning
- support vector
- text classification
- learning algorithm
- transfer learning
- statistical analysis
- learning process
- clustering algorithm
- sample size
- high quality
- state space