Prediction algorithms and confidence measures based on algorithmic randomness theory

被引：49

作者：

Gammerman, A ^{[1
]}

Vovk, V

机构：

[1] Univ London Royal Holloway & Bedford New Coll, Comp Learning Res Ctr, Egham TW20 0EX, Surrey, England

[2] Univ London Royal Holloway & Bedford New Coll, Dept Comp Sci, Egham TW20 0EX, Surrey, England

来源：

THEORETICAL COMPUTER SCIENCE | 2002年 / 287卷 / 01期

基金：

英国工程与自然科学研究理事会;

关键词：

confidence; machine learning; randomness; transduction;

D O I：

10.1016/S0304-3975(02)00100-7

中图分类号：

TP301 [理论、方法];

学科分类号：

081202 ;

摘要：

This paper reviews some theoretical and experimental developments in building computable approximations of Kolmogorov's algorithmic notion of randomness. Based on these approximations a new set of machine learning algorithms have been developed that can be used not just to make predictions but also to estimate the confidence under the usual iid assumption. (C) 2002 Elsevier Science B.V. All rights reserved.

引用

页码：209 / 217

页数：9

共 12 条

[1]

[Anonymous], 1989, P ADV NEURAL INFORM

[2]

GAMMERMAN A, 1998, P 14 C UNC ART INT M, P148

[3]

LI M, 1997, INTRO KOLMOGOROV COM

[4]

Melluish Thomas, 2001, TYPICALNESS FRAMEWOR

[5]

Nouretdinov I, 2001, LECT NOTES ARTIF INT, V2111, P337

[6]

NOURETDINOV I, 2001, P 18 INT C MACH LEAR

[7]

PROEDROU K, 2002, P ART INT STAT C

[8]

Saunders C, 1999, IJCAI-99: PROCEEDINGS OF THE SIXTEENTH INTERNATIONAL JOINT CONFERENCE ON ARTIFICIAL INTELLIGENCE, VOLS 1 & 2, P722

[9]

Scholkopf B., 1998, Advances in Kernel Methods-Support Vector Learning

[10]

Vapnik V., 1998, STAT LEARNING THEORY, V1, P2

← 1 2 →