Tight conditions for consistent variable selection in high dimensional nonparametric regression.
Laëtitia CommingesArnak S. DalalyanPublished in: COLT (2011)
Keyphrases
- variable selection
- high dimensional
- nonparametric regression
- input variables
- cross validation
- linear models
- low dimensional
- stochastic search
- lower bound
- high dimensionality
- model selection
- high dimensional data
- group lasso
- data points
- dimension reduction
- feature space
- worst case
- dimensionality reduction
- upper bound
- nearest neighbor
- supervised learning
- sparsity inducing
- number of input variables
- linear combination
- pointwise
- sparse representation
- machine learning
- reinforcement learning
- bayesian networks