Feature Selection for Mutlti-labeled Variables via Dependency Maximization.
Salimeh Yasaei SekehAlfred O. Hero IIIPublished in: ICASSP (2019)
Keyphrases
- feature selection
- variable selection
- dependency structure
- markov blanket
- text categorization
- mutual information
- feature selection algorithms
- text classification
- information gain
- machine learning
- objective function
- support vector
- classification accuracy
- multi class
- training data
- method for feature selection
- manual labeling
- conditional dependencies
- random variables
- unsupervised learning
- multi task
- supervised learning
- semantic dependencies
- feature space
- model selection
- input variables
- small sample
- support vector machine
- knn
- redundant features
- probabilistic model
- bayesian networks