Measures for the evaluation and comparison of graphical model structures

Research output: Chapter in Book/Report/Conference proceedingsConference contributionpeer-review

2 Citations (Scopus)

Abstract

Structure learning is the identification of the structure of graphical models based solely on observational data and is NP-hard. An important component of many structure learning algorithms are heuristics or bounds to reduce the size of the search space. We argue that variable relevance rankings that can be easily calculated for many standard regression models can be used to improve the efficiency of structure learning algorithms. In this contribution, we describe measures that can be used to evaluate the quality of variable relevance rankings, especially the well-known normalized discounted cumulative gain (NDCG). We evaluate and compare different regression methods using the proposed measures and a set of linear and non-linear benchmark problems.

Original languageEnglish
Title of host publicationComputer Aided Systems Theory – EUROCAST 2017 - 16th International Conference, Revised Selected Papers
EditorsRoberto Moreno-Diaz, Alexis Quesada-Arencibia, Franz Pichler
PublisherSpringer Verlag
Pages283-290
Number of pages8
ISBN (Print)9783319747170
DOIs
Publication statusPublished - 2018
Event16th International Conference on Computer Aided Systems Theory, EUROCAST 2017 - Las Palmas de Gran Canaria, Spain
Duration: 19 Feb 201724 Feb 2017

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume10671 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference16th International Conference on Computer Aided Systems Theory, EUROCAST 2017
CountrySpain
CityLas Palmas de Gran Canaria
Period19.02.201724.02.2017

Keywords

  • Graphical models
  • Regression
  • Structure learning

Fingerprint Dive into the research topics of 'Measures for the evaluation and comparison of graphical model structures'. Together they form a unique fingerprint.

Cite this