PAC-Bayesian risk bounds for group-analysis sparse regression by exponential weighting

Comparing our proposed globally sparse PLS with standard PLS and L 1 penalized PLS [Chun et al., 2010], we see that PLS with global variable selection attains better performance

The non- linear bayesian kernel regression can therefore be consid- ered as achieved online by a Sigma Point Kalman Filter.. First experiments on a cardinal sine regression show

To this aim we propose a parsimonious and adaptive decomposition of the coefficient function as a step function, and a model including a prior distribution that we name

When the mixed norms are used in the framework of regularized inverse problem, a thresholded Landweber iteration is used to minimize the corresponding variational problem.. Key

Theorem 2.1 provides an asymptotic result for the ridge estimator while Theorem 2.2 gives a non asymptotic risk bound of the empirical risk minimizer, which is complementary to

The aim of this paper is to provide the first accel- eration of exponential weights procedures achieving the optimal rate of convergence d 0 log(d)/(αT ) in the identically

We study the problem of detection of a p-dimensional sparse vector of parameters in the linear regression model with Gaussian noise.. We establish the detection boundary, i.e.,

Our main theoretical result is a sparsity oracle inequality in probability for the true excess risk for a version of exponential weight estimator.. We also propose a MCMC method