Combining localized fusion and dynamic selection for high-performance SVM. Issue 1 (January 2015)

• Record Type: Journal Article
• Authors: Min, Jun-Ki; Hong, Jin-Hyuk; Cho, Sung-Bae
• Is Part Of: Expert systems with applications. Volume 42:Issue 1(2015)
• Page Start: 9
• Page End: 20
• Publication Date: 2015-01
• Subjects: Expert systems (Computer science) -- Periodicals; Systèmes experts (Informatique) -- Périodiques; Electronic journals; 006.33
• DOI: 10.1016/j.eswa.2014.07.028
• Languages: English
• ISSNs: 0957-4174

Abstract

<abstract xml:lang="en" abstract-type="author-highlights" id="ab015"> <title id="st075">Highlights</title> <sec> <p id="sp0095"> <list id="l0005"> <list-item id="o0005"> <label></label> <p id="p0225">We present multiple SVMs combined by localized fusion and dynamic selection.</p> </list-item> <list-item id="o0010"> <label></label> <p id="p0230">Clustering OVA SVMs trained for each class constructs multiple decision templates.</p> </list-item> <list-item id="o0015"> <label></label> <p id="p0235">A naïve Bayes classifier selects the localized models dynamically for a new sample.</p> </list-item> <list-item id="o0020"> <label></label> <p id="p0240">Experiments on nine benchmark datasets show its superiority against alternatives.</p> </list-item> <list-item id="o0025"> <label></label> <p id="p0245">It proves to effectively manage the unbiased-variance and bias in real datasets.</p> </list-item> </list> </p> </sec> </abstract>