Active learning for clinical text classification: is it better than random sampling?
Rosa L. FigueroaQing Zeng-TreitlerLong H. NgoSergey GoryachevEduardo P. WiechmannPublished in: J. Am. Medical Informatics Assoc. (2012)
Keyphrases
- random sampling
- text classification
- active learning
- machine learning
- selective sampling
- query by committee
- text categorization
- uncertainty sampling
- feature selection
- labeled data
- sampling algorithm
- supervised learning
- naive bayes
- text mining
- sampling methods
- stratified sampling
- sampling procedure
- unlabeled data
- adaptive sampling
- learning algorithm
- random samples
- random sample
- random projections
- semi supervised learning
- semi supervised
- training set
- multi label
- transfer learning
- knn
- reservoir sampling
- sample size
- training examples
- learning process
- decision trees
- class imbalance
- cost sensitive
- data cleaning
- generalization error
- natural language processing
- upper bound
- data structure
- data sets