Title :
A new principle for measuring the generalization performance of SVMs
Author :
Zhou, Weida ; Zhang, Li ; Jiao, Licheng
Author_Institution :
Nat. Key Lab. for Radar Signal Process., Xidian Univ., Xi´´an, China
Abstract :
A new method for estimating the VC dimension in advance is proposed, which can be taken as a principle for measuring the generalization performance of SVM. Our method adopts the two-order statistic of training samples and maintains consistency with the method for estimating the VC dimension in statistical learning theory. Our method can be applied to the problem of model selection and sample pre-processing. Simulation results show the feasibility and practicability of our method.
Keywords :
generalisation (artificial intelligence); learning automata; learning by example; sampling methods; SVM; VC dimension estimation; generalization performance measurement; model selection; sample pre-processing; training samples; two-order statistic; Erbium; Neural networks; Pattern recognition; Performance evaluation; Radar signal processing; Signal processing algorithms; Statistical learning; Statistics; Support vector machines; Virtual colonoscopy;
Conference_Titel :
Signal Processing, 2002 6th International Conference on
Print_ISBN :
0-7803-7488-6
DOI :
10.1109/ICOSP.2002.1179989
