Investigating classifier learning behavior with experiment databases

J. Vanschoren, Hendrik Blockeel

Research output: Chapter in Book/Report/Conference proceedingConference contributionAcademicpeer-review

3 Citations (Scopus)


Experimental assessment of the performance of classification algorithms is an important aspect of their development and application on real-world problems. To facilitate this analysis, large numbers of such experiments can be stored in an organized manner and in complete detail in an experiment database. Such databases serve as a detailed log of previously performed experiments and a repository of verifiable learning experiments that can be reused by different researchers. We present an existing database containing 250,000 runs of classifier learning systems, and show how it can be queried and mined to answer a wide range of questions on learning behavior. We believe such databases may become a valuable resource for classification researchers and practitioners alike.
Original languageEnglish
Title of host publicationData analysis, machine learning and applications
Subtitle of host publicationProceedings of the 31st Annual Conference of the Gesellschaft für Klassifikation e.V., Albert-Ludwigs-Universität Freiburg, March 7–9, 2007
EditorsChristine Preisach, Hans Burkhardt, Lars Schmidt-Thieme
Place of PublicationBerlin
Number of pages8
ISBN (Electronic)978-3-540-78246-9
ISBN (Print)978-3-540-78239-1
Publication statusPublished - 2008
Externally publishedYes

Publication series

NameStudies in Classification, Data Analysis, and Knowledge Organization (STUDIES CLASS)


Dive into the research topics of 'Investigating classifier learning behavior with experiment databases'. Together they form a unique fingerprint.

Cite this