Online Stream Deviator
A module that alters the data from incoming streams in order to hide the precise values while preserving statistical properties of the data.
Date Posted: April 19, 2007
|
|
 |
|
|
What is Online Stream Deviator?
Online Stream Deviator is a module that alters the data from incoming streams in order to hide the precise values while preserving statistical properties of the data. This technology can be used by streaming data dissemination applications where the consumers are interested in analyzing trends, finding clusters, and discovering similarities and correlations across multiple streams and where the publisher does not want to reveal the real values of sensitive data.
A complete description of the data perturbation algorithm used in Online Stream Deviator
can be found in the following paper: Hiding in the Crowd: Privacy Preservation on Evolving Streams through Correlation Tracking, by Feifei Li, Jimeng Sun, Spiros Papadimitriou, George Mihaila, and Ioana Stanoi. Proceedings of the 2007 IEEE 23rd International Conference on Data Engineering (ICDE 2007), Istanbul, Turkey, April 2007.
How does it work? Online Stream Deviator continuously tracks the inter-stream correlations and introduces noise exhibiting the same correlations, thus hampering the efforts to recover the real values by exploiting correlations.
|
|
|
| | About the technology author(s):
Feifei Li, a Ph.D. candidate, does research that covers indexing and query processing in spatio-temporal database, algorithms, and applications in stream and sensor databases. Mr. Li's recent research interests focus on security and privacy issues in both relational database and stream database systems. More specifically, he is working on privacy preservation of the data as well as authentication and verification of query results.
Jimeng Sun, a Ph.D. candidate, is primarily interested in the fields of data mining, machine learning, and database systems. Mr. Sun's research focuses on the design, analysis, implementation, and experimental evaluation of streaming data-mining algorithms.
Spiros Papadimitriou, Ph.D., is a research staff member at IBM® Watson Research Center. His main interests are in time series, data mining on streaming data, clustering, and privacy. He has published more than 20 papers on these topics in refereed conferences and journals. Dr. Papadimitriou was a recipient of a Siebel scholarship.
George Mihaila, Ph.D., is a research staff member at IBM Watson Research Center and he holds an adjunct faculty appointment at Columbia University. Dr. Mihaila's research interests include Web query
languages, Web-based information discovery, data integration, data warehousing, event processing, and XML storage and processing.
Ioana Stanoi, Ph.D., is a research staff member at IBM Almaden Research Center. Her patents and publications cover exact and approximate query processing, index optimization, XML, publish/subscribe systems, mobile clients, e-commerce applications, and semantics.
IBM is a trademark of IBM Corporation in the United States, other countries, or both.
Other company, product, or service names may be trademarks or service marks of others.
|
|
|
| |
