APPrOVE: Application-oriented Validation and Evaluation of Supervised Learners

Document type: Conference Papers
Peer reviewed: Yes
Full text:
Author(s): Niklas Lavesson, Paul Davidsson
Title: APPrOVE: Application-oriented Validation and Evaluation of Supervised Learners
Conference name: IEEE Intelligent Systems
Year: 2010
Pagination: 150-155
ISBN: 978-1-4244-5164-7
Publisher: IEEE press
City: London
Organization: Blekinge Institute of Technology
Department: School of Computing (Sektionen för datavetenskap och kommunikation)
School of Computing S-371 79 Karlskrona
+46 455 38 50 00
Authors e-mail:,
Language: English
Abstract: Learning algorithm evaluation is usually focused on classification performance. However, the characteristics and requirements of real-world applications vary greatly. Thus, for a particular application, some evaluation criteria are more important than others. In fact, multiple criteria need to be considered to capture application-specific trade-offs. Many multi-criteria methods can be used for the actual evaluation but the problems of selecting appropriate criteria and metrics as well as capturing the trade-offs still persist. This paper presents a framework for application-oriented validation and evaluation (APPrOVE). The framework includes four sequential steps that together address the aforementioned problems and its use in practice is demonstrated through a case study.
Subject: Computer Science\Artificial Intelligence
Software Engineering\General
Keywords: classification, evaluation, supervised learning