2009-2010 University Scholar Profile
Steven Buss
Mentor: Tim Davis
College of Engineering
"I applied in order to receive both monetary support and, more importantly, faculty assistance. The conversations I have had with my mentor and other faculty members have been invaluable. Create a scalable recommender system using advanced machine learning techniques on large, sparse datasets which are constantly changing."
Courses of Study
Major
Computer Science
Minor
Mathematics
Research Interests
Data mining, machine learning, and robotics
Academic Awards
- Dean's List
- Bright Futures Scholar
Hobbies/Activities
Gardening, cooking, and programming
Research Description
Connectious: A Scalable, Adaptive, Domain-Agnostic Recommender System
Connectious is a recommender system that learns user preferences automatically and continuously via several machine learning techniques. It then uses this knowledge to produce relevant results for a user. After an extensive literature review, a few techniques were selected for implementation and possible improvement. A key criterion for selection was scalability, with parallelization a primary goal. These techniques were implemented with Hadoop, a distributed processing framework for large scale computation. This research has many real-world applications, including the website that is part of the research project, http://www.connectious.com which showcases the recommendation technology in the domain of web site discovery. Future applications include business intelligence and organization-specific information discovery and data categorization.
Back to the Journal of Undergraduate Research
