Shapley Values for Feature Selection: The Good, the Bad, and the Axioms.
Daniel Vidali FryerInga StrümkeHien D. NguyenPublished in: IEEE Access (2021)
Keyphrases
- feature selection
- information gain
- axiomatic characterization
- attribute values
- knowledge base
- feature space
- discriminative features
- feature selection algorithms
- text categorization
- unsupervised feature selection
- small sample
- parameter values
- first order logic
- mutual information
- feature set
- machine learning
- neural network
- text classification
- standard deviation
- dimensionality reduction
- knowledge representation
- training data
- irrelevant features
- decision trees