PAC analyses of a ‘similarity learning’ IBL algorithm

  • A. D. Griffiths
  • , D. G. Bridge

Research output: Chapter in Book/Report/Conference proceedingsConference proceedingpeer-review

Abstract

VS-CBR [14] is a simple instance-based learning algorithm that adjusts a weighted similarity measure as well as collecting cases. This paper presents a ‘PAC’ analysis of VS-CBR, motivated by the PAC learning framework, which demonstrates two main ideas relevant to the study of instance-based learners. Firstly, the hypothesis spaces of a learner on different target concepts can be compared to predict the difficulty of the target concepts for the learner. Secondly, it is helpful to consider the ‘constituent parts’ of an instance-based learner: to explore separately how many examples are needed to infer a good similarity measure and how many examples are needed for the case base. Applying these approaches, we show that VS-CBR learns quickly if most of the variables in the representation are irrelevant to the target concept and more slowly if there are more relevant variables. The paper relates this overall behaviour to the behaviour of the constituent parts of VS-CBR.

Original languageEnglish
Title of host publicationCase-Based Reasoning Research and Development - 2nd International Conference on Case-Based Reasoning, ICCBR 1997, Proceedings
EditorsEnric Plaza, David B. Leake
PublisherSpringer Verlag
Pages445-454
Number of pages10
ISBN (Print)3540632336, 9783540632337
DOIs
Publication statusPublished - 1997
Externally publishedYes
Event2nd International Conference on Case-Based Reasoning, ICCBR 1997 - Providence, United States
Duration: 25 Jul 199727 Jul 1997

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume1266
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference2nd International Conference on Case-Based Reasoning, ICCBR 1997
Country/TerritoryUnited States
CityProvidence
Period25/07/9727/07/97

Fingerprint

Dive into the research topics of 'PAC analyses of a ‘similarity learning’ IBL algorithm'. Together they form a unique fingerprint.

Cite this