<論文・報告>Efficient Filtering Algorithm with Fused Constraint

Abstract

Sparse modeling is an effective machine-learning method to analyze missing and sparse data. The fused LASSO (least absolute shrinkage and selection operator) is one of the techniques used in sparse modeling. In this paper, we propose a simple yet effective optimization technique for the fused LASSO problem based on the method [1]. We then show the main advantage of the proposed technique: the proposed update formula monotonically decreases the objective function.