A knowledge-rich similarity measure for improving IT incident resolution process
2010 (English)In: Proceedings of the 25th Annual ACM Symposium on Applied Computing 2010: Sierre, Switzerland, March 22 - 26, 2010, New York: ACM Digital Library, 2010, 1781-1788 p.Conference paper (Refereed)
The aim of incident management is to restore a given IT service disruption, simply called incident, to normal state as quickly as possible. In incident management, it is essential to resolve a new incident efficiently and accurately. However, typically, incident resolution process is largely manual, thus, it is time-consuming and error-prone. This paper proposes a new knowledge-rich similarity measure for improving this process. The role of this measure is to retrieve the most similar past incident cases for a new incident without human intervention. The solution information contained the retrieved incident cases can be utilized to resolve the new incident. The main feature of our similarity measure is to incorporate additional useful meta knowledge, outside of incident description that is the only exploited information in typical similarity measures used in CBR, to improve effectiveness. Moreover, this measure exploits as much semantic knowledge as possible about features contained in previous incident cases. Through an experimental evaluation, we show the effectiveness, technical coherence and feasibility of this measure using a real dataset
Place, publisher, year, edition, pages
New York: ACM Digital Library, 2010. 1781-1788 p.
IdentifiersURN: urn:nbn:se:ltu:diva-34110DOI: 10.1145/1774088.1774466Local ID: 837eea50-e7f9-11df-8b36-000ea68e967bISBN: 978-1-60558-639-7 (print)OAI: oai:DiVA.org:ltu-34110DiVA: diva2:1007360
ACM Symposium on Applied Computing : 22/03/2010 - 26/03/2010
Upprättat; 2010; 20101104 (andbra)2016-09-302016-09-30