Department of Mathematics

Colloquium

  •  An Auction Dynamics Approach to Semi-Supervised Data Classification
  •  04/17/2018
  •  4:10 PM - 5:00 PM
  •  C304 Wells Hall
  •  Ekaterina Rapinchuk, Michigan State University

We reinterpret the semi-supervised data classification problem using an auction dynamics framework inspired by real life auctions. This novel forward and reverse auction procedure for data classification requires remarkably little training/labeled data and readily incorporates volume/class size constraints. We prove that the algorithm always terminates with the right properties for any choice of parameters and derive its computational complexity. Experimental results on benchmark machine learning datasets show that our approach exceeds the performance of current state-of-the-art methods, while requiring a fraction of the computational time. This is joint work with Matt Jacobs and Selim Esedoglu.

 

Contact

Department of Mathematics
Michigan State University
619 Red Cedar Road
C212 Wells Hall
East Lansing, MI 48824

Phone: (517) 353-0844
Fax: (517) 432-1562

College of Natural Science