|
I am a PhD candidate in the Database Lab in
the Computer Science Department at Indiana University. My research interests are
DB theory, data mining, and heuristic algorithms. For my thesis I am developing the MoLS framework,
which is a customizable tool for mining Approximate Functional Dependencies (AFDs). The development of
MoLS was based on taking a software engineering approach to define a conceptual modularization of algorithms
use to search powerset lattices. Customizing the different elements of algorithms are done through plug-ins.
The focus of developing MoLS is to create a testbed for designing customized algorithms and exploring how
and algorithm can be customized for better performance. In this pursuit we have developed the AK-DFS algorithm
which improves performance over previous AFD mining algorithms by orders of magnitude. Another research interest both as part of the MoLS project and outside projects is the development of heuristic guided algorithms.
|
