Embed Size (px)
Citation preview
Martin [email protected]
OntoNaviERP: Ontology-supported Navigation in ERP Software
Documentation
Martin Hepp1,2 and Andreas Wechselberger1
1E-Business and Web Science Research Group, Bundeswehr University Munich, Germany
2STI Innsbruck, Austria
{[email protected] | [email protected]}
1
Martin [email protected]
ERP: Enterprise Resource PlanningMRP: Material Requirements Planning
MRP-II: Manufacturing Resource Planning
CIM: Computer Integrated Manufacturing
2
Comprehensive --> Complex
Martin [email protected]
Search is important, because ...
• Users must align their mental models of processes and data with those underlying the ERP software
• Intensive usage
• More efficient usage of ERP documentation means more efficient business operations.
4
?Function
Event
Function
Document
Role
Martin [email protected]
Search is difficult, because ...
5
Textbook Business Terminology
SAP Business Terminology
SAP IT/System Terminology
Industry Domain Terminology
Martin [email protected]
Requirements
• Cost-efficient Ontology Construction
• Cost-efficient Annotation
• Speed
• User Interface
7
Martin [email protected]
Which Terms Are Most Valuable - The Most Popular, the Frequent, or the Rare?
Frequency vs. Information Value vs. Likelihood of Relevance
10
Web-w
ide N
um
ber
of O
ccurr
ences
Term 1..N (similar pattern for: Tags, Ontology Elements)
+ Likelihood of relevant documents
+ Likelihood of usage in queries
- limited information value
balance between frequency and specificity
+ higher information value
+ highest information value if needed
- likelihood of occurrence in query low
- likelihood of relevant documents low
Martin [email protected]
Approach: Conceptual Model
11
Semi-automatically derived from
term lists
Traditional ontology
engineering
Martin [email protected]
Approach: Ontology Construction (1)
• Script
• Protégé
• WordNet Plug-In
• OWL Annotation Property
• .NT Generator
12
Martin [email protected]
Results (1)
19
Median of retrieved documents• Term-based: 0
>=50% of the action-topic pairs yield no result• OntoNaviERP: 5.5
50% of of the action-topic pairs return at least 5 result pages, of which 2 are relevant
Martin [email protected]
Findings and Lessons Learned
• Understanding of ontology engineering and annotation under economic constraints
• In our setting, consolidation of synonyms and lexical variants does a great deal of the job
21
Martin [email protected]
Future Extensions
• Task and Context
• Annotate individual sections of resources
• Skills / Users’ Competencies vs. Target Audience of a Resource
• Hybrid Search: Combining term-based and conceptual searchE.g. see Bhagdev/Chapman/Ciravegna, Vitaveska Lanfranchi, Daniela Petrelli: Hybrid Search: Effectively Combining Keywords and Semantic Searches, ESWC 2008
22
Martin [email protected]
Reference and LinksHepp, Martin and Wechselberger, Andreas: OntoNaviERP: Ontology-supported Navigation in ERP Software Documentation, in: A. Sheth et al. (eds.): ISWC 2008. Proceedings of the 7th International Semantic Web Conference, Germany, October 26-30, 2008, Karlsruhe, Germany, Springer LNCS, Vol. 5318, pp. 764-776.
This paper and other papers are available at
http://www.heppnetz.de/publications/
The .NT Generator for KIM/GATE Gazetteer Lists is at:
http://ebusiness-unibw.org:8180/ntfilegenerator/
23
Martin [email protected]
Thank you!Martin Hepp and Andreas Wechselberger
{[email protected] | [email protected]}
http://www.unibw.de/ebusiness/
24
![Intro ERP Using GBI Navigation course[Letter] en v2.01 ERP 6.0 EhP4 Navigation Author Stefan Weidner to put their knowledge into practice Version 2.01 MOTIVATION This material explains](https://img.dokumen.tips/doc/110x75/5aba74e17f8b9a567c8b8abb/intro-erp-using-gbi-navigation-courseletter-en-v201-erp-60-ehp4-navigation-author.jpg)
