Why-oriented end-user debugging of naive Bayes text classification.
Todd KuleszaSimone StumpfWeng-Keen WongMargaret M. BurnettStephen PeronaAmy J. KoIan OberstPublished in: ACM Trans. Interact. Intell. Syst. (2011)
Keyphrases
- end users
- text classification
- naive bayes
- text categorization
- logistic regression
- naive bayes classifier
- text mining
- feature selection
- user interface
- uci datasets
- cost sensitive
- bag of words
- bayesian classifier
- uci data sets
- text classifiers
- text documents
- machine learning
- classification algorithm
- co training
- text data
- averaged one dependence estimators
- bayesian network classifiers
- semantic features
- knn
- labeled data
- unlabeled data
- test instances
- document classification
- multi label
- naive bayes classification
- locally weighted
- term frequency
- artificial intelligence
- base classifiers
- information retrieval
- bayesian classifiers
- naive bayesian classifier
- probabilistic classifiers
- attribute dependencies
- k nearest neighbor
- data mining
- training data
- conditional independence assumption
- pairwise