Efficient Coverage of Case Space with Active Learning

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

Authors

Nuno Filipe Escudeiro (Author) Other; Alipio Mario Jorge (Author) FCUP

Conference proceedings International

Title: Progress in Artificial Intelligence: Proceedings of the 14th Portuguese Conference on Artificial Intelligence, EPIA2009

Pages: 411-422

14th EPIA - Portuguese Conference on Artificial Intelligence - EPIA 2009

Universidade de Aveiro, Portugal, 12 a 15 de Outubro de 2009

Indexing

ISI Web of Knowledge - 2 Citations

Scopus - 3 Citations

Scientific classification

FOS: Natural sciences > Computer and information sciences

Other information

Authenticus ID: P-003-RW2

DOI: 10.1007/978-3-642-04686-5_34

Abstract (EN): Collecting and annotating exemplary cases is a costly and critical task that is required in early stages of any classification process. Reducing labeling cost without degrading accuracy calls for a compromise solution which may be achieved with active learning. Common active learning approaches focus on accuracy and assume the availability of a pre-labeled set of exemplary cases covering all classes to learn. This assumption does not necessarily hold. In this paper we study the capabilities of a new active learning approach, d-Confidence, in rapidly covering the case space when compared to the traditional active learning confidence criterion, when the representativeness assumption is not met.. Experimental results also show that; d-Confidence reduces the number of queries required to achieve complete class coverage and tends to improve or maintain classification error.

Language: English

Type (Professor's evaluation): Scientific

No. of pages: 12