DocumentCode :
2179533
Title :
Binary Classification Using Linear SVM Pyramidal Tree
Author :
Kumar, Arun M. ; Gopal, M.
Author_Institution :
C&O Res., ABB Global Ind. & Services Ltd., Bangalore, India
fYear :
2010
fDate :
9-10 Feb. 2010
Firstpage :
54
Lastpage :
58
Abstract :
This paper presents a linear SVM (Support Vector Machine) Pyramidal Tree (SVMPT) for binary classification tasks. SVMPT is a modified version of SVM based Tree Type Neural Networks (SVMTNN), reported earlier in the literature [1]. Both the algorithms use parameter-less SVM proposed by Mangasarian [2] for learning in each node. While SVMTNN insists on 100 percent training accuracy, linear SVMPT uses predetermined threshold value to determine when to stop adding new nodes. Experimental results on standard binary datasets show that the algorithm has good generalization capability, comparable to linear SVMs. We also present experimental results on extensions of linear SVMPT to multi-class datasets and datasets with fuzzy membership for each datapoint.
Keywords :
fuzzy set theory; pattern classification; support vector machines; trees (mathematics); SVM based tree type neural networks; binary classification; fuzzy membership; linear SVM pyramidal tree; multiclass datasets; support vector machine; Classification tree analysis; Delay; Feedforward neural networks; Multi-layer neural network; Network topology; Neural networks; Neurons; Robustness; Support vector machine classification; Support vector machines; Fuzzy SVM; Pattern classification; SVM tree; support vector machines;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Data Storage and Data Engineering (DSDE), 2010 International Conference on
Conference_Location :
Bangalore
Print_ISBN :
978-1-4244-5678-9
Type :
conf
DOI :
10.1109/DSDE.2010.31
Filename :
5452638
Link To Document :
بازگشت