Presentation	2003/9/9 Stepwise Entropy Analysis Kenichi Yoshida,
PDF Download Page	PDF download Page Link
Abstract(in Japanese)	(See Japanese page)
Abstract(in English)	The importance of analyzing semi-structured data, such as hyper linked WWW text, XML data and chemical formulae, promotes research on a group of data-mining methods which can handle trees, graphs, and other non-table format data. Among these, GBI: Graph Based Induction [1,2] is one of the incipient methods to extract hidden rules from graph-format data. This paper describes an algorithm, SEA: Stepwise Entropy Analysis, using the GBI method. Though the conventional GBI algorithm contracts with the input graph during the analysis, this new algorithm does not use a contraction operation. Although this new algorithm is still greedy, its greediness is slightly weakened by omitting the contraction operation. Another characteristic of SEA, the simultaneous discovery of the association rule and the classification rule, is also described.
Keyword(in Japanese)	(See Japanese page)
Keyword(in English)	Graph based induction / entropy
Paper #	AI2003-62
Date of Issue

Conference Information
Committee	AI
Conference Date	2003/9/9(1days)
Place (in Japanese)	(See Japanese page)
Place (in English)
Topics (in Japanese)	(See Japanese page)
Topics (in English)
Chair
Vice Chair
Secretary
Assistant

Paper Information
Registration To	Artificial Intelligence and Knowledge-Based Processing (AI)
Language	ENG
Title (in Japanese)	(See Japanese page)
Sub Title (in Japanese)	(See Japanese page)
Title (in English)	Stepwise Entropy Analysis
Sub Title (in English)
Keyword(1)	Graph based induction
Keyword(2)	entropy
1st Author's Name	Kenichi Yoshida
1st Author's Affiliation	Graduate School of Business Science, University of Tsukuba()
Date	2003/9/9
Paper #	AI2003-62
Volume (vol)	vol.103
Number (no)	306
Page	pp.pp.-
#Pages	5
Date of Issue