Adaptive algorithms in sampling recovery

Abstract

We study optimal algorithms in adaptive sampling recovery of smooth functions defined on the unit $d$-cube ${\II}^d:= [0,1]^d$. The recovery error is measured in the quasi-norm $\|\cdot\|_q$ of $L_q := L_q(\II^d)$. For $B$ a subset in $L_q,$ we define a sampling algorithm of recovery with the free choice of sample points and recovering functions from $B$ as follows. For each $f$ from the quasi-normed Besov space $B^\alpha_{p,\theta}$, we choose $n$ sample points. This choice defines $n$ sampled values. Based on these sample points and sampled values, we choose a function from $B$ for recovering $f$. The choice of $n$ sample points and a recovering function from $B$ for each $f \in B^\alpha_{p,\theta}$ defines a $n$-sampling algorithm $S_n^B$ by functions in $B$. If $\Phi = \{\phi_k\}_{k \in K}$ is a family of elements in $L_q$, let $\Sigma_n(\Phi)$ be the non-linear set of linear combinations of $n$ free terms from $\Phi,$ that is $\Sigma_n(\Phi):= \{\, \phi = \sum_{j=1}^n a_j \phi_{k_j}: \ k_j \in K \, \}$. Denote by ${\mathcal G}$ the set of all families $\Phi$ in $L_q$ such that the intersection of $\Phi$ with any finite dimensional subspace in $L_q$ is a finite set, and by $\Cc(B^\alpha_{p,\theta}, L_q)$ the set of all continuous mappings from $B^\alpha_{p,\theta}$ into $L_q$. We define the quantity $$\nu_n(B^\alpha_{p,\theta},L_q) := \inf_{\Phi \in {\mathcal G}} \inf_{S_n^B \in \Cc(X, L_q): B= \Sigma_n(\Phi)} \sup_{\|f\|_{B^\alpha_{p,\theta}} \le 1} \ \|f - S_n^B(f)\|_q.$$ Let $0 < p,q, \theta \le \infty $ and $\alpha > d/p$. Then we prove the asymptotic order $$ \nu_n(B^\alpha_{p,\theta},L_q) \asymp n^{- \alpha / d}.$$ We also obtained the asymptotic order of quantities of optimal recovery by $S_n^B$ in terms of best $n$-term approximation as well of other non-linear $n$-widths.