|Dept. of Computer Sciences
University of Wisconsin - Madison
1210 West Dayton Street
Madison, WI 53706-1685
Kevin S. Beyer
Advisor: Raghu Ramakrishnan
Areas of Interest: Database Systems, Data Analysis, OLAP, Visualization, High Dimensional Data, Performance, Data Mining, ...
Thesis: Interactive Data Analysis using Query Set Templates
I am investigating techniques to support families of related queries that arise during data analysis. My ideas were inspired by (and will be implemented in) the DEVise project. In DEVise, many views of data are simultaneously presented to the analyst. The data needs for each view is represented by one query template (i.e., a parameterized query), and the entire collection of queries is called a query set template, or Qset for short. Once the data needs are described by a Qset, several opportunities arise. In particular, I am interested in multiple query optimization and processing, automatic index selection, and automatic materialized view selection.
My other main interest is in extending relational database systems with support for sequence data. A relation is an unordered collection of records. Often, however, there is an underlying order (e.g., a sequence of stock prices), and users want to pose queries that reflect this order (e.g., find a weekly moving average). SQL provides little support for posing such queries. My colleagues and I have proposed an extension to SQL, called SRQL, that can express a rich class of queries over sorted relations. We believe SRQL can be easily and efficiently supported in current RDBMSs with only minor extensions.
"Nearest Neighbor"Meaningful? Kevin S. Beyer, Jonathan Goldstein, Raghu Ramakrishnan, Uri Shaft. In Proceedings of the 7th International Conference on Database Theory (ICDT), January 1999. (ps.gz)