• Aucun résultat trouvé

Oracle inequalities, aggregation and adaptation

N/A
N/A
Protected

Academic year: 2021

Partager "Oracle inequalities, aggregation and adaptation"

Copied!
161
0
0

Texte intégral

Loading

Figure

Figure 1 is a particular case of the general mirror averaging algorithm of Iouditski et al
Figure 2. Linear Mirror Averaging algorithm (lma).
Figure 1. Sample splitting scheme
Table 1. MISE for the Gaussian (left) and the exponential (right) densities
+7

Références

Documents relatifs

The main contributions of this result is to provide more flexibility on the choices of j 0 and j 1 (which will be crucial in our dependent framework) and to clarified the

To reach this aim, numerous inequalities exist: Markov’s inequality, Tchebychev’s in- equality, Chernoff’s inequality, Berry-Esseen’s inequality, Bernstein’s inequality,

corresponds to the rate of convergence obtained in the estimation of f (m) in the 1-periodic white noise convolution model with an adapted hard thresholding wavelet estimator

Building on very elegant but not very well known results by George [31], they established sharp oracle inequalities for the exponentially weighted aggregate (EWA) for

Huang and Zhang [17] considered the general Group Lasso setting but obtained only bounds for prediction and ℓ 2 estimation errors, while [22] focused only on the multi-task

there is a decrease in the quality of working conditions combined with an increase in the intensities of technical and market constraints and a decrease in

Using an aggregate with exponential weights, we obtain an optimal rate of convex aggregation for the hinge risk under the margin assumption.. Moreover, we obtain an optimal rate

In brief, such techniques were based on the analytic continuation of a conformal mapping, requiring some a priori topological knowledge of the free boundary and