We describe novel fast learning curve method - methods for scaling inductive methods to large data sets - and their application to clustering. We describe the decision theoretic underpinnings of the approach and demonstrate significant performance gains on two real-world data sets.

PDF file

In  Proceedings of Eighth International Workshop on Artificial Intelligence and Statistics

Publisher  Morgan Kaufmann Publishers
All copyrights reserved by Morgan Kaufmann Publishers 2007.

Details