2006/012 - Categorisation in multiple category systems
Gabriela Csurka, Eric Gaussier, Cyril Goutte, Francois Pacull, Jean-Michel Renders
International conference on machine learning, (ICML) Pittsburgh, Pennsylvania, June 25-29, 2006.
We explore the situation in which documents have to be categorised into more than one category system, a situation we refer to as multiple-view categorisation. Such a situation arises for example in large companies where incoming mails have to be routed to several departments, each one relying on its own category system. We focus here on exploiting possible dependencies between category systems in order to refine the categorisation decisions made by categorisers trained independently on different category systems. After a description of the multiple categorisation problem, we present several possbile solutions, based either on a categorisation or reweighting approach, and compare them on real data. Lastly, we show how the multi-media categorisation problem can be cast as a multiple categorisation problem and assess our methods in this framework