Description: Description: Description: Worcester Polytechnic Institute (WPI)

 

 

Description: Description: Description: http://web.cs.wpi.edu/images/cs_banner.gif

             COLLOQUIUM

 
Complex Event Analytics: 
From Real-time Business Intelligence Services to Healthcare Applications
[Or, on how this sabbatical went by like a short blip.]
 

 

Elke Angelika Rundensteiner

Professor, Computer Science Department

Worcester Polytechnic Institute

rundenst@cs.wpi.edu

Sabbatical, what sabbatical? As always time is too short, as there are too many interesting problems to work on.  In this presentation, I'll briefly introduce some of my on-going collaborative research projects

in the area of "Complex Event Stream Analytics", with applications in Real-time Business Intelligence Services (HP) and in the Healthcare domain (UMASS).

 

Common to modern applications from infection control tracking to RFID-based supply chain management systems is that they transmit real-time data streams. Thus, there is a pressing need for event stream processing technology to analyze this vast amount of sequential data to enable online operational decision making. However, state-of-the-art Complex Event Processing (CEP) systems designed for sequence detection neither support traditional online analytical processing operations nor powerful actions.

 

First, I'll introduce our recent results on the E-Cube model – an approach of combining CEP and OLAP techniques for efficient multi-dimensional event pattern analysis at different abstraction levels (SIGMOD'2011).  This E-Cube research, produced in collaboration with students and faculty at WPI and researchers from Hewlett-Packard Labs, CA, is being supported by a National Science Foundations grant and an HP Innovations grant.  Given a workload of CEP queries, based on interrelationships in both concept abstraction and pattern refinement among these queries, E-Cube composes the queries into an integrated event pattern query hierarchy. Based on this hierarchy, strategies of drill-down and roll-up are supported.  A cost-driven optimizer for optimal E-Cube workload execution is also introduced. Experimental studies comparing alternate strategies on a real world financial data stream demonstrate the superiority of the proposed E-Cube technology.

 

Second, I'll touch upon results from our collaborative project with students at WPI and researchers from UMass Medical School supported by an NSF grant and a UMASS/WPI collaborative grant.  Event Processing technology, while effective for pattern query execution, is limited in its capability of reacting to risks detected by pattern queries. Reactions that affect the query results in turn have not been

addressed in the current state-of-art. I'll describe our solution to this problem (published in VLDB'2011), namely, to embed active rule support within the CEP engine, henceforth called Active CEP (ACEP).

Active rules in ACEP allow us to specify a pattern query's dynamic condition and real-time actions.  Our solution to the technical challenge of handling interactions between queries and reactions for queries in the high-volume stream execution are also sketched.  We demonstrate the power of the ACEP technology as solution for supporting business logic in real-time event-based systems – by applying it as a case study to the development of a healthcare system currently deployed in UMass Medical School hospital.

___________

 

Elke A. Rundensteiner is a Full Professor in the Computer Science Department of Worcester Polytechnic Institute (WPI). Elke received her B.S. degree (Vordiplom) from the Johann Wolfgang Goethe University, Frankfurt, West Germany, an Master's degree from the Florida State University, Tallahassee, and a Ph.D. degree from the University of California, Irvine, all in Computer Science.

Prof. Rundensteiner is an internationally recognized expert in databases and information systems, having spend 20 years of her career focussing on the development of scalable data management technology in support of advanced applications including business, engineering, and sciences.  Her current research interests include scalable data stream processing, query optimization, complex event analytics, information integration and visual exploration, and data warehousing for distributed systems.  She has over 300 publications in these and related areas.  Her publications on view technology, database integration, and data evolution are widely cited, and her research software prototypes released to public domain have been used by academic and non-profit groups around the world.  Her research has been funded by government agencies including NSF, NIH, DOE and by industry and government labs including HP, IBM, Verizon Labs, GTE, NEC, Mitre Corporation, Charles River Analytics, and others.

She has been recipient of numerous honors, including NSF Young Investigator, Sigma Xi Outstanding Senior Faculty Researcher, and WPI Board of Trustees' Outstanding Research and Creative Scholarship award, and the 2010 WPI Chairman's Exemplary Faculty Prize.  She is on program committees of prestigious conferences in the database field, has been editor of several journals, including Associate Editor of the IEEE Transactions on Data and Knowledge Engineering Journal, and of the VLDB Journal, and serving as PC chair of several conferences, currently of EDBT'2012. Prof. Rundensteiner runs the database group DSRG.

 

____

 

Refreshments will be served at 10:45 a.m. in Fuller Labs, third floor –lounge area.

 

 

 

 

 

 

 

 

 

 

Description: Description: Description: [WPI]Description: Description: Description: [Home]Description: Description: Description: [Top]