Convex body domination and weighted estimates with matrix weights (Q2401691): Difference between revisions

Let \(f \in L^1(Q)\) with \(Q \subset \mathbb{R}^{d}\) and \(\Im\) be a sparse family of cubes. We define its convex body averages \(\ll f\gg_Q\) by \[ \ll f \gg_Q := \{ <\varphi f>_Q \, \mid \, \varphi : Q \rightarrow \mathbb{R}, \;\|\varphi\|_{\infty} \leq 1 \}, \] and a sparse (Lerner) operator \(\mathbf{L} = \mathbf{L}_{\Im}\) by \[ \mathbf{L}_{\Im}f = \sum_{Q \in \Im} \ll f \gg_Q \mathbf{1}_Q, \] where the sum is Minkowsky sum. By a simple computation, in order to estimate a convex-body valued sparse operator \(\mathbf{L}_{\Im}\) one needs to find a uniform bound on all operators of the form \[ f \rightarrow \sum_{Q \in \Im} \int K_Q(x,y) f(y) \, dy \] where kernels \(K_Q\) are supported on \(Q \times Q\) and \(\|K_Q\|_{\infty} \leq |Q|^{-1}\). In this paper, the authors introduce a notion of domination of vector valued singular integrals by sparse operators. They firstly show that if a scalar operator \(T\) can be dominated by a sparse operator, then its vector version of the form \(T \otimes \mathbf{1}_d\) can be dominated by a convex body valued sparse operator \(\mathbf{L}_{\Im}\). They especially proved that \(\omega\)-Calderón-Zygmund operators with the modulus of continuity satisfying the Dini condition, and for a wide class of martingale operators, including the so-called big Haar shifts and paraproducts are dominated by sparse operators, where \(\omega\) is an increasing subadditive function on \([0, \infty)\) with \(\omega(0)=0\) and \(\|\omega\|_{\text{Dini}}:= \int_{0}^{1} \omega(x) \frac{dx}{x} < \infty\). They also obtain some \(A_2\)-\(A_{\infty}\) type weighted estimates of vector valued operators \(T\) for matrix weights even in a two weight setting, where \(T\) is either an \(\omega\)-Calderón-Zygmund operator or a big Haar shift or a paraproduct.