Improving Text Classification Using EM with Background Text.
Sarah ZelikovitzHaym HirshPublished in: FLAIRS Conference (2005)
Keyphrases
- text classification
- text data
- text documents
- text mining
- text classifiers
- document categorization
- unsupervised learning
- bag of words
- training corpus
- text categorization
- text representation
- naive bayes
- expectation maximization
- feature selection
- labeled data
- em algorithm
- multi label
- document classification
- textual data
- machine learning
- n gram
- web documents
- maximum likelihood
- text retrieval
- complex background
- automatic text classification
- sentiment classification
- foreground objects
- sentiment analysis
- probabilistic model
- image segmentation
- semantic features
- text collections
- k means
- data cleaning
- text information
- association rules
- decision trees
- information retrieval