Change search
ReferencesLink to record
Permanent link

Direct link
An ADMM Algorithm for a Class of Total Variation Regularized Estimation Problems
KTH, School of Electrical Engineering (EES), Centres, ACCESS Linnaeus Centre. (Automatic Control)ORCID iD: 0000-0002-1927-1690
Stanford University.
KTH, School of Electrical Engineering (EES), Automatic Control. KTH, School of Electrical Engineering (EES), Centres, ACCESS Linnaeus Centre. (System Identification Group)
Stanford University.
2012 (English)In: Preprints of the 16th IFAC Symposium on System Identification / [ed] Kinnaert, Michel, 2012, 83-88 p.Conference paper, Presentation (Refereed)
Abstract [en]

We present an alternating augmented Lagrangian method for convex optimization problems where the cost function is the sum of two terms, one that is separable in the variable blocks, and a second that is separable in the difference between consecutive variable blocks. Examples of such problems include Fused Lasso estimation, total variation denoising, and multi-period portfolio optimization with transaction costs. In each iteration of our method, the first step involves separately optimizing over each variable block, which can be carried out in parallel. The second step is not separable in the variables, but can be carried out very efficiently. We apply the algorithm to segmentation of data based on changes in mean (l_1 mean filtering) or changes in variance (l_1 variance filtering). In a numerical example, we show that our implementation is around 10000 times faster compared with the generic optimization solver SDPT3.

Place, publisher, year, edition, pages
2012. 83-88 p.
Keyword [en]
Machine Learning and Data Mining; Maximum Likelihood Methods; Time Series
National Category
Control Engineering
URN: urn:nbn:se:kth:diva-101909DOI: 10.3182/20120711-3-BE-2027.00310ScopusID: 2-s2.0-84867091929OAI: diva2:549832
16th IFAC Symposium on System Identification, Square - Brussels Meeting Centre, Belgium
ICT - The Next Generation

QC 20121112

Available from: 2012-11-12 Created: 2012-09-05 Last updated: 2013-09-05Bibliographically approved

Open Access in DiVA

0310(330 kB)418 downloads
File information
File name FULLTEXT01.pdfFile size 330 kBChecksum SHA-512
Type fulltextMimetype application/pdf

Other links

Publisher's full textScopusIFAC-papersonline

Search in DiVA

By author/editor
Wahlberg, BoAnnergren, Mariette
By organisation
ACCESS Linnaeus CentreAutomatic Control
Control Engineering

Search outside of DiVA

GoogleGoogle Scholar
Total: 418 downloads
The number of downloads is the sum of all downloads of full texts. It may include eg previous versions that are now no longer available

Altmetric score

Total: 105 hits
ReferencesLink to record
Permanent link

Direct link