Sparse Group Fused Lasso for Model Segmentation
share
This article introduces the sparse group fused lasso (SGFL) as a statistical framework for segmenting high dimensional regression models. To compute solutions of the SGFL, a nonsmooth and nonseparable convex program, we develop a hybrid optimization method that is fast, requires no tuning parameter selection, and is guaranteed to converge to a global minimizer. In numerical experiments, the hybrid method compares favorably to state-of-the-art techniques both in terms of computation time and accuracy; benefits are particularly substantial in high dimension. The hybrid method is implemented in the R package sparseGFL available on the author's Github page. The SGFL framework, presented here in the context of multivariate time series, can be extended to multichannel images and data collected over graphs.
READ FULL TEXT
