Title
An open architecture for event-based analytics.
Abstract
Event-based analytics is increasingly gaining prominence in business and social applications. Despite the availability of many solutions specializing in event processing systems (e.g. CEP technology), there is currently no commonly agreed way of describing event and event pattern types, and thus no standardized method for interchange of event pattern instances between systems. This paper advocates an open architecture for event-based analytics comprising a common model that supports interoperability of data between different systems. It introduces the foundational concepts for describing event patterns including events, event pattern types and event pattern occurrences. The event pattern meta-model is also formalized using a UML meta-model to facilitate its adoption and usage across the event analytics community. The paper provides a case study introducing several event pattern types from the financial market data analytics domain. This case study illustrates a number of specific event pattern types used by finance experts and an application that requires interoperability between two separate software component frameworks (a rule-based front-end and a CEP). Results show that the meta-model concepts are sufficient to represent and implement a class of real-life business analytics solutions. The paper also identified a number of semantic challenges in developing interoperability solutions for the event-based processing, in spite of the fact that we needed to merge only two separately developed event-based conceptual models.
Year
DOI
Venue
2016
10.1007/s41060-016-0029-7
I. J. Data Science and Analytics
Keywords
Field
DocType
Complex event processing, Event-based analytics, Real-time analytics, Data modelling, Service-oriented architecture, ADAGE
Data science,Business analytics,Software analytics,Computer science,Interoperability,Complex event processing,Semantic analytics,Event correlation,Analytics,Service-oriented architecture
Journal
Volume
Issue
ISSN
2
1-2
2364-4168
Citations 
PageRank 
References 
1
0.41
7
Authors
4
Name
Order
Citations
PageRank
Zoran Milosevic154854.38
Weisi Chen212.10
Andrew Berry310.41
Fethi Rabhi442750.68