A creative abduction approach to scientific and knowledge discovery

H. Prendinger, M. Ishizuka

Research output: Contribution to journalArticle

12 Citations (Scopus)

Abstract

In this paper we explore the range of applicability of abductive reasoning for knowledge discovery. In particular, we discuss a novel form of abduction, called creative abduction, where new knowledge is generated in the process of explaining observed events, and demonstrate its relevance to knowledge discovery. The main contribution of this paper is twofold: First, we show that creative abduction can be used to infer a disposition explaining local temporal regularities. In the presence of multiple correlated regularities, this form abduction may significantly unify a given corpus of knowledge, corresponding to theory formation in scientific discovery. Second, we present a weaker form of creative abduction that infers a goal (e.g. interest) from simple 'condition-effect' rules called 'transitions'. If multiple transitions are correlated, the weaker form of creative abduction can be used to identify, e.g. clusters of Web users, as done in Web usage mining. We will focus on the formal underpinnings of this new form of abduction that seems readily applicable to a wide range of practical knowledge discovery problems.

Original languageEnglish
Pages (from-to)321-326
Number of pages6
JournalKnowledge-Based Systems
Volume18
Issue number7
DOIs
Publication statusPublished - 2005 Nov
Externally publishedYes

Fingerprint

Data mining
Abduction
Knowledge discovery

Keywords

  • Abduction
  • Logic for knowledge discovery
  • Scientific discovery

ASJC Scopus subject areas

  • Artificial Intelligence

Cite this

A creative abduction approach to scientific and knowledge discovery. / Prendinger, H.; Ishizuka, M.

In: Knowledge-Based Systems, Vol. 18, No. 7, 11.2005, p. 321-326.

Research output: Contribution to journalArticle

Prendinger, H. ; Ishizuka, M. / A creative abduction approach to scientific and knowledge discovery. In: Knowledge-Based Systems. 2005 ; Vol. 18, No. 7. pp. 321-326.
@article{a9eb44f7ee674ec8b3a1c547acbc8fc9,
title = "A creative abduction approach to scientific and knowledge discovery",
abstract = "In this paper we explore the range of applicability of abductive reasoning for knowledge discovery. In particular, we discuss a novel form of abduction, called creative abduction, where new knowledge is generated in the process of explaining observed events, and demonstrate its relevance to knowledge discovery. The main contribution of this paper is twofold: First, we show that creative abduction can be used to infer a disposition explaining local temporal regularities. In the presence of multiple correlated regularities, this form abduction may significantly unify a given corpus of knowledge, corresponding to theory formation in scientific discovery. Second, we present a weaker form of creative abduction that infers a goal (e.g. interest) from simple 'condition-effect' rules called 'transitions'. If multiple transitions are correlated, the weaker form of creative abduction can be used to identify, e.g. clusters of Web users, as done in Web usage mining. We will focus on the formal underpinnings of this new form of abduction that seems readily applicable to a wide range of practical knowledge discovery problems.",
keywords = "Abduction, Logic for knowledge discovery, Scientific discovery",
author = "H. Prendinger and M. Ishizuka",
year = "2005",
month = "11",
doi = "10.1016/j.knosys.2004.12.003",
language = "English",
volume = "18",
pages = "321--326",
journal = "Knowledge-Based Systems",
issn = "0950-7051",
publisher = "Elsevier",
number = "7",

}

TY - JOUR

T1 - A creative abduction approach to scientific and knowledge discovery

AU - Prendinger, H.

AU - Ishizuka, M.

PY - 2005/11

Y1 - 2005/11

N2 - In this paper we explore the range of applicability of abductive reasoning for knowledge discovery. In particular, we discuss a novel form of abduction, called creative abduction, where new knowledge is generated in the process of explaining observed events, and demonstrate its relevance to knowledge discovery. The main contribution of this paper is twofold: First, we show that creative abduction can be used to infer a disposition explaining local temporal regularities. In the presence of multiple correlated regularities, this form abduction may significantly unify a given corpus of knowledge, corresponding to theory formation in scientific discovery. Second, we present a weaker form of creative abduction that infers a goal (e.g. interest) from simple 'condition-effect' rules called 'transitions'. If multiple transitions are correlated, the weaker form of creative abduction can be used to identify, e.g. clusters of Web users, as done in Web usage mining. We will focus on the formal underpinnings of this new form of abduction that seems readily applicable to a wide range of practical knowledge discovery problems.

AB - In this paper we explore the range of applicability of abductive reasoning for knowledge discovery. In particular, we discuss a novel form of abduction, called creative abduction, where new knowledge is generated in the process of explaining observed events, and demonstrate its relevance to knowledge discovery. The main contribution of this paper is twofold: First, we show that creative abduction can be used to infer a disposition explaining local temporal regularities. In the presence of multiple correlated regularities, this form abduction may significantly unify a given corpus of knowledge, corresponding to theory formation in scientific discovery. Second, we present a weaker form of creative abduction that infers a goal (e.g. interest) from simple 'condition-effect' rules called 'transitions'. If multiple transitions are correlated, the weaker form of creative abduction can be used to identify, e.g. clusters of Web users, as done in Web usage mining. We will focus on the formal underpinnings of this new form of abduction that seems readily applicable to a wide range of practical knowledge discovery problems.

KW - Abduction

KW - Logic for knowledge discovery

KW - Scientific discovery

UR - http://www.scopus.com/inward/record.url?scp=30144445559&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=30144445559&partnerID=8YFLogxK

U2 - 10.1016/j.knosys.2004.12.003

DO - 10.1016/j.knosys.2004.12.003

M3 - Article

AN - SCOPUS:30144445559

VL - 18

SP - 321

EP - 326

JO - Knowledge-Based Systems

JF - Knowledge-Based Systems

SN - 0950-7051

IS - 7

ER -