Differentiable Feature Selection by Discrete Relaxation.
Rishit ShethNicoló FusiPublished in: AISTATS (2020)
Keyphrases
- feature selection
- objective function
- feature space
- text categorization
- support vector
- machine learning
- mutual information
- discrete version
- high dimensionality
- discrete geometry
- classification accuracy
- feature selection algorithms
- iterative algorithms
- multi task
- support vector classification
- irrelevant features
- redundant features
- continuous relaxation
- information gain
- feature set
- dimensionality reduction
- support vector machine
- knn
- feature extraction
- microarray data
- loss function
- feature weighting
- model selection
- unsupervised feature selection
- forward selection
- data sets