Research on tangent circular arc smooth Support Vector Machine (TCA-SSVM) algorithm

Data classification problem is a flourishing research field. Classification is the process of finding the common properties among different patterns and classifying them into classes. SVM (support vector machine) is one of classifier for solving binary classification problem. In traditional SVM solu...

Full description

Saved in:
Bibliographic Details
Published in2008 International Conference on Information and Automation pp. 1322 - 1327
Main Authors Yan-Feng Fan, De-Xian Zhang
Format Conference Proceeding
LanguageEnglish
Published IEEE 01.06.2008
Subjects
Classification
Classification algorithms
Data mining
DFP Quasi-Newton
Kernel
Optimization
Polynomials
smooth technology
Smoothing methods
Support vector machines
SVM
tangent circular arc polynomial
Online AccessGet full text

Cover

More Information
Summary:Data classification problem is a flourishing research field. Classification is the process of finding the common properties among different patterns and classifying them into classes. SVM (support vector machine) is one of classifier for solving binary classification problem. In traditional SVM solution algorithms, objective function is a strictly convex unconstrained optimization problem, but is un-differentiable due to plus function x + , which precludes the most used optimization algorithms. A new smoothing technology which replaces the plus function by an accurate tangent circular arc polynomial for solving SVM classification algorithm is proposed in this paper. We also prescribe a DFP quasi-Newton algorithm to solve the proposed classifier. Numerical results and comparisons are given to demonstrate the effectiveness.
ISBN:1424421837
9781424421831
DOI:10.1109/ICINFA.2008.4608206