UCI plus plus : Improved Support for Algorithm Selection Using Datasetoids

• Type: Article in International Conference Proceedings Book
• Date: 2009

Authors

Carlos Soares (Author) FEP

Conference proceedings International

Title: Advances in Knowledge Discovery and Data Mining

Pages: 499-506

13th Pacific-Asia Conference on Knowledge Discovery and Data Mining (PAKDD 2009)

Bangkok, Thailand, April 27-30, 2009

Indexing

ISI Web of Knowledge - 17 Citations

Scopus - 22 Citations

Scientific classification

FOS: Natural sciences > Computer and information sciences

Other information

Authenticus ID: P-003-NXP

DOI: 10.1007/978-3-642-01307-2_46

Abstract (EN): As companies employ a larger number of models, the problem of algorithm (and parameter) selection is becoming increasingly important. Two approaches to obtain empirical knowledge that is useful for that purpose are empirical studies and metalearning. However, most empirical (meta)knowledge is obtained from a, relatively small set, of datasets. In this paper, we propose a method to obtain a large number of datasets which is based on a simple transformation of existing datasets, referred to as datasetoids. We test our approach on the problem of using metalearning to predict when to prune decision trees. The results show significant; improvement when using datasetoids. Additionally, we identify a number of potential anomalies in the generated datasetoids and propose methods to solve them.

Language: English

Type (Professor's evaluation): Scientific

No. of pages: 8