Enabling user-driven rule management in event data analysis

• Published in: Information systems frontiers Vol. 18; no. 3; pp. 511 - 528
• Main Authors: Chen, Weisi, Rabhi, Fethi A.
• Format: Journal Article
• Language: English
• Published: New York Springer US 01.06.2016; Springer Nature B.V
• Subjects: Analysis; Business and Management; Computer programs; Control; Data analysis; Data processing; Evolution; Information systems; Information technology; Infrastructure; IT in Business; Learning; Management; Management of Computing and Information Systems; Operations Research/Decision Theory; Organization theory; Pattern recognition; Platforms; Small & medium sized enterprises-SME; Social networks; Software; Studies; Subject specialists; Systems Theory; Event processing; Event meta-model; Rule management; Ripple down rules; Event data
• ISSN: 1387-3326; 1572-9419
• DOI: 10.1007/s10796-016-9633-2

Event data analysis is becoming increasingly of interest to academic researchers looking for patterns in the data. Unlike domain experts working in large companies who have access to IT staff and expensive software infrastructures, researchers find it harder to efficiently manage their event data analysis by themselves. Particularly, user-driven rule management is a challenge especially when analysis rules increase in size and complexity over time. In this paper, we propose an event data analysis platform called EP-RDR intended for non-IT experts that facilitates the evolution of event processing rules according to changing requirements. This platform integrates a rule learning framework called Ripple-Down Rules (RDR) operating in conjunction with an event pattern detection component invoked as a service (EPDaaS). We have built a prototype to demonstrate this solution on real-life scenario involving financial data analysis.