A Novel Vectors Surgeon Machine Based on Statistical Learning Theory

Yu LIANG, Liang-zhi GAN

Abstract


Support vector machine (SVM) was first used to solve classification problems and then developed to solve regression problems. Despite its widespread success, the SVM suffers from some important limitations, one of the most significant being that its optimization algorithms to train the SVM are complicated. We investigate the problem of training SVM and present a novel machine learning method that is called Vectors Surgeon Regression Machine (VSRM). The VSMR minimizes the squares error and controls the structural risk by the number of parameters, which is equal to the VC dimension of the set of functions. The optimal brain surgeon usually used in neural networks is introduced to prune the support vectors. The proposed method is tested on classification data sets.

Keywords


Support vector machine, Optimal brain surgeon, Statistical learning theory

Publication Date


2016-12-21 00:00:00


DOI
10.12783/dteees/seeie2016/4555

Full Text:

PDF

Refbacks

  • There are currently no refbacks.