Event-driven Architecture for Health Event Detection from Multiple - - PowerPoint PPT Presentation

1

Event-driven Architecture for Health Event Detection from Multiple Sources

• Dr. Kerstin Denecke

Oslo, 31.08.2011

Agenda

• Motivation
• Event-driven Architectures
• Architecture for Health Event Detection
• Conclusions and Future Work

8 September 2011 Kerstin Denecke 2

Motivation: The Threat

• Infectious diseases lead to public health

threats

• Contributing factors:

n global warming, globalization, air traffic …mutation, spread, …

• Technologies to early detection are

required è Early Detection by Disease Surveillance è Broad range of sources

8 September 2011 Kerstin Denecke 3

Requirements for (improved) surveillance

• Diversity of information sources
• (Additional) information in time
• Filtering along relevance, novelty, source...
• Visualization with innovative techniques
• User feedback and interaction

è Real-time processing of huge amounts of data

8 September 2011 Kerstin Denecke 4

Event-driven Architectures

8 September 2011 Kerstin Denecke 5

• Behaviour of a system is orchestrated around the

production, detection and consumption of events

• Information can be propagated in near-real-time

throughout a highly distributed environment

• Event:
• a significant or meaningful change in state
• message, token, count, pattern
• based on an asynchronous message-driven

communication mode

System architecture

8 September 2011 Kerstin Denecke 6

User Modeling and Recommendation Indicator Detection and Signal Generation User interface Content Collector and Document Analysis TV / Radio Twitter Blogs Online News

8 September 2011 Kerstin Denecke 7

Information flow

M-Eco System – Result Page

8 September 2011 Kerstin Denecke 8

Conclusions and Future Directions

• Surveillance system that follows event-driven

design principle

• Currently first prototype is tested
• First observations:

n difficulties in collecting the right data n algorithms are sometimes slow when processing large data sets

è Improvements are necessary (data stream processing)

8 September 2011 Kerstin Denecke 9

References

• Denecke K et al.: The Medical Ecosystem - Personalised Event-based
• Surveilllance. In: Safran C, Reti S, Marin HF: Studies in Health Technology and

Informatics: MEDINFO 2010. Proceedings of the 13th World Congress on Medical and Health Informatics, IOS Press, Amsterdam, 2010, ISBN 978-1-60750-587-7, pp. 1382

• Stewart A, Fisichella M, Denecke K: Detecting Public Health Indicators from the

Web for Epidemic Intelligence. 3rd International Conference on Electronic Healthcare (eHealth 2010), December 13-15, 2010, Casablanca, Morocco

• Fisichella M, Stewart A, Denecke K, Nejdl W: Unsupervised Public Health Event

Detection for Epidemic Intelligence. CIKM’10, October 25-29, 2010, Toronto, Ontario, Canada

• Denecke K, Stewart A: Learning from Medical Social Media Data: Current State

and Future Challenges. In: White B, King I, Tsang P: Social Computing in Higher Learning Environments: How Media, Tools, and Platforms Change Learner’s

• Behaviour. Springer Verlag, Heidelberg, March 2011

8 September 2011 Kerstin Denecke 10

Thank you for your attention!

• Dr. Kerstin Denecke

Forschungszentrum L3S denecke@L3S.de http://www.meco-project.eu

Event-Driven Architectures for Health Event Detection from Multiple Sources