Title :
A modified ν-SV method for simplified regression
Author :
Shilton, A. ; Palaniswami, M.
Author_Institution :
Dept. of Electr. & Electron. Eng., Melbourne Univ., Parkville, Vic., Australia
Abstract :
In the present paper we describe a new algorithm for support vector regression (SVR). Like standard ν-SVR algorithms, this algorithm automatically adjusts the radius of insensitivity (tube width ε) to fit the data. However, this is achieved without additional complexity in the optimisation problem. Moreover, by careful modification of the kernel function, we are able to significantly simplify the form of the dual SVR optimisation problem.
Keywords :
computational complexity; optimisation; regression analysis; support vector machines; computational complexity; kernel function; modified ν-algorithm; optimisation; support vector regression; Australia; Cost function; Integrated circuit noise; Kernel; Neural networks; Noise reduction; Pattern recognition; Support vector machines; Training data;
Conference_Titel :
Intelligent Sensing and Information Processing, 2004. Proceedings of International Conference on
Print_ISBN :
0-7803-8243-9
DOI :
10.1109/ICISIP.2004.1287694
