We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:


Current browse context:


Change to browse by:

References & Citations

DBLP - CS Bibliography


(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo ScienceWISE logo

Computer Science > Machine Learning

Title: Fused Lasso for Feature Selection using Structural Information

Abstract: Feature selection has been proven a powerful preprocessing step for high-dimensional data analysis. However, most state-of-the-art methods tend to overlook the structural correlation information between pairwise samples, which may encapsulate useful information for refining the performance of feature selection. Moreover, they usually consider candidate feature relevancy equivalent to selected feature relevancy, and some less relevant features may be misinterpreted as salient features. To overcome these issues, we propose a new feature selection method using structural correlation between pairwise samples. Our idea is based on converting the original vectorial features into structure-based feature graph representations to incorporate structural relationship between samples, and defining a new evaluation measure to compute the joint significance of pairwise feature combinations in relation to the target feature graph. Furthermore, we formulate the corresponding feature subset selection problem into a least square regression model associated with a fused lasso regularizer to simultaneously maximize the joint relevancy and minimize the redundancy of the selected features. To effectively solve the optimization problem, an iterative algorithm is developed to identify the most discriminative features. Experiments demonstrate the effectiveness of the proposed approach.
Subjects: Machine Learning (cs.LG); Machine Learning (stat.ML)
Cite as: arXiv:1902.09947 [cs.LG]
  (or arXiv:1902.09947v3 [cs.LG] for this version)

Submission history

From: Lu Bai [view email]
[v1] Tue, 26 Feb 2019 14:17:06 GMT (244kb)
[v2] Tue, 25 Jun 2019 19:52:23 GMT (4471kb,D)
[v3] Fri, 28 Jun 2019 19:30:14 GMT (261kb)

Link back to: arXiv, form interface, contact.