A probabilistic framework for SVM regression and error bar estimation

被引：154

作者：

Gao, JB ^{[1
]}

Gunn, SR

Harris, CJ

Brown, M

机构：

[1] Univ Southampton, Dept Elect & Comp Sci, Image Speech & Intelligent Syst Res Grp, Southampton SO17 1BJ, Hants, England

[2] IBM Hursley Lab, HS&T, Data Explos Grp, Winchester SO21 2JN, Hants, England

来源：

MACHINE LEARNING | 2002年 / 46卷 / 1-3期

基金：

中国国家自然科学基金;

关键词：

support vector machine (SVM); Gaussian process; epsilon-loss function; error bar estimation;

D O I：

10.1023/A:1012494009640

中图分类号：

TP18 [人工智能理论];

学科分类号：

081104 [模式识别与智能系统]; 0812 [计算机科学与技术]; 0835 [软件工程]; 1405 [智能科学与技术];

摘要：

In this paper, we elaborate on the well-known relationship between Gaussian Processes (GP) and Support Vector Machines (SVM) under some convex assumptions for the loss functions. This paper concentrates on the derivation of the evidence and error bar approximation for regression problems. An error bar formula is derived based on the epsilon -insensitive loss function.

引用

页码：71 / 89

页数：19

共 25 条

[1]

NEW LOOK AT STATISTICAL-MODEL IDENTIFICATION [J].

AKAIKE, H .

IEEE TRANSACTIONS ON AUTOMATIC CONTROL, 1974, AC19 (06) :716-723

[2]

[Anonymous], 1987, P CBMS NSF REG C SER

[3]

EVGENIOU T, 1999, 1654 MIT AI LAB

[4]

GAO J, 2000, VARIATIONAL APPROACH

[5]

An equivalence between sparse approximation and support vector machines [J].

Girosi, F .

NEURAL COMPUTATION, 1998, 10 (06) :1455-1480

[6]

Gunn S. R., 1998, SUPPORT VECTOR MACHI

[7]

Gunn SR, 1997, LECT NOTES COMPUT SC, V1280, P313

[8]

Moderating the outputs of support vector machine classifiers [J].

Kwok, JTY .

IEEE TRANSACTIONS ON NEURAL NETWORKS, 1999, 10 (05) :1018-1031

[9]

MacKay D.J., 1997, GAUSSIAN PROCESSES R

[10]

A PRACTICAL BAYESIAN FRAMEWORK FOR BACKPROPAGATION NETWORKS [J].

MACKAY, DJC .

NEURAL COMPUTATION, 1992, 4 (03) :448-472

← 1 2 3 →