Feature Selection Using Improved Mutual Information for Text Classification.
Jana NovovicováAntonín MalíkPavel PudilPublished in: SSPR/SPR (2004)
Keyphrases
- mutual information
- feature selection
- text classification
- information theoretic
- information gain
- text categorization
- conditional mutual information
- naive bayes
- bag of words
- n gram
- labeled data
- medical image registration
- feature selection algorithms
- machine learning
- text mining
- classification accuracy
- multi class
- support vector machine
- feature subset
- information theoretic criterion
- multimodal image registration
- feature reduction
- irrelevant features
- image registration
- feature weighting
- semantic features
- multiresolution
- data cleaning
- support vector
- text data
- feature set
- text documents
- knn
- unlabeled data
- unsupervised learning
- k nearest neighbor
- model selection
- term frequency
- text classifiers
- multi label
- similarity measure
- neural network
- distributional clustering