Information Systems
Dr. George Karabatis
Associate Professor

Department of Information Systems, UMBC
George Karabatis


Home
Announcements
Research
Education
Publications
Teaching
Professional Activities



Below are some topics for students interested in independent studies. If interested contact Dr. George Karabatis, georgek AT umbc DOT edu to set up an appointment and discuss further details.

Independent Study on Information Integration
This independent study focuses on a very important problem in private/public organizations, government, or anyplace that needs access to data: How to integrate information stored in different databases. Although the simple extraction of data from multiple databases followed by simple union of the results has been used extensively in daily database applications, it has been proven that this approach is insufficient with today’s vast amounts of semantically varying data.
This Independent Study targets this exact research problem, and students are expected to apply semantic techniques to databases to provide semantically meaningful data integration with an emphasis on semantic network techniques. The mode of the course will be dual: A research problem with literature review, and a proposed implementation solution.
Requirements: Students should be knowledgeable of, or willing to learn advanced database concepts.
Application domains: Environmental databases, or Biological databases depending on student preference.

 

Independent Study on Data Quality
This Independent Study deals with a real problem that has been in existence since the dawn of databases. The huge amount of data in databases is going ‘stale’. The data is also getting ‘dirty’ and queries to the databases do not bring back results consistent with reality. Many errors occur at different phases of the lifecycle of data, resulting in ‘dirty’ data. This independent study targets this problem. Specifically, the following questions will be addressed:
Why data get dirty? What are the root causes? What is the percentage of dirty data? How to create mechanisms to alleviate this problem? Now I updated and corrected the dirty data. How to guarantee that it will stay current?
The mode of the course will be dual: A research problem with literature review and a proposed implementation solution.
Requirements: Students should be knowledgeable of advanced database concepts.
Application domains: Environmental databases, or Biological databases depending on student preference.