Multivariate Trend Filtering for Lattice Data

arXiv:2112.14758v2 Announce Type: replace-cross
Abstract: We study a multivariate version of trend filtering, called Kronecker trend filtering or KTF, for the case in which the design points form a lattice in $d$ dimensions. KTF is a natural extension of univariate trend filtering (Steidl et al., 2006; Kim et al., 2009; Tibshirani, 2014), and is defined by minimizing a penalized least squares problem whose penalty term sums the absolute (higher-order) differences of the parameter to be estimated along each of the …

abstract arxiv case cs.lg data design dimensions extension filtering form lattice math.st multivariate natural stat.ml stat.th study trend type