Local rotation-based ensemble. (25th August 2010)

• Record Type: Journal Article
• Authors: Kotsiantis, S.B.
• Is Part Of: International journal of knowledge engineering and data mining. Volume 1:Number 2(2010)
• Page Start: 147
• Page End: 160
• Publication Date: 2010-08-25
• Subjects: data mining -- supervised machine learning -- classification -- local ensemble -- rotation-based ensemble -- weak classifiers -- rotation forests -- local regions -- data characteristics -- target class; Knowledge representation (Information theory) -- Periodicals; Data mining -- Periodicals; 006.305
• Languages: English
• ISSNs: 1755-2087

Abstract

Many data analysis problems involve an investigation of relationships between attributes in heterogeneous databases, where different prediction models can be more appropriate for different regions. We propose a technique of local rotation-based ensemble of weak classifiers. In order to determine rotation forests, we identify local regions having similar characteristics and then build local classification experts on each of these regions describing the relationship between the data characteristics and the target class. We performed a comparison with other well-known combining methods using weak classifiers as based learners, on standard benchmark datasets and we took better accuracy.