Leave-one-out cross-validation is risk consistent for lasso.
Darren HomrighausenDaniel J. McDonaldPublished in: Mach. Learn. (2014)
Keyphrases
- leave one out cross validation
- variable selection
- model selection
- cross validation
- feature selection
- ridge regression
- generalization error
- support vector machine
- k nearest neighbour
- high dimensional
- confidence intervals
- sample size
- dimension reduction
- data sets
- least squares
- kernel learning
- machine learning
- regression model
- linear combination
- feature subset
- linear regression
- knn
- evolutionary algorithm
- ls svm
- selection criterion
- training set
- feature extraction