MIKM: A mutual information-based K-medoids approach for feature selection.
Jung-Yi JiangYao-Lung SuShie-Jue LeePublished in: ICMLC (2011)
Keyphrases
- feature selection
- mutual information
- image registration
- information gain
- conditional mutual information
- feature selection algorithms
- feature set
- text categorization
- machine learning
- text classification
- irrelevant features
- support vector machine
- classification accuracy
- feature ranking
- small sample
- selected features
- feature weighting
- unsupervised feature selection
- multi task
- model selection
- data sets
- multi class
- feature extraction
- naive bayes
- feature space
- support vector
- bayes classifier
- informative features
- decision trees
- redundant features
- computer vision
- spatial clustering with obstacles constraints