BS-SVM multi-classification model in the application of consumer goods

• Published in: 2011 IEEE 2nd International Conference on Software Engineering and Service Science pp. 1 - 4
• Main Authors: Quanhui Jia, Lieli Liu
• Format: Conference Proceeding
• Language: English
• Published: IEEE 01.07.2011
• Subjects: Binary Search Tree; Classification algorithms; Information systems; Injuries; multi-classification; Relative Separability; Rough sets; Safety; support vector machine; Support vector machines; Training
• ISBN: 9781424496990; 1424496993
• ISSN: 2327-0586
• DOI: 10.1109/ICSESS.2011.5982240

Quality and safety of consumer products have drawn wide attention from scholars in related domain, this issue is based on the subject of the quality and safety of consumer goods, in accordance with characteristics of cases, and put forward a hierarchical support vector machine classification algorithm based on the relative separability of the feature space, to solve the low classification performance and high rate of misclassification of the existing algorithms. The weight of Binary Search Tree is the separability of samples, determining the order of categories by a selective set of training samples to construct SVM classifier and the final formation of a binary classification of the larger interval multi-valued SVM classifier tree. Simulation results show that the method has a faster test speed, relatively perfect good classification accuracy and generalization performance.