Binarsity: a penalization for one-hot encoded features in linear supervised learning.
Mokhtar Z. AlayaSimon BussyStéphane GaïffasAgathe GuillouxPublished in: J. Mach. Learn. Res. (2019)
Keyphrases
- supervised learning
- training examples
- feature set
- feature space
- feature detection
- variable selection
- spatial information
- unsupervised learning
- co occurrence
- low level
- feature vectors
- generalization error
- feature extraction
- search engine
- machine learning
- specific features
- closed form
- extracted features
- benchmark datasets
- image features
- prior knowledge
- training set
- reinforcement learning
- training data
- learning algorithm