Benelearn'08: The Annual Belgian-Dutch Machine Learning Conference pages:63-64
Benelearn'08: : The Annual Belgian-Dutch Machine Learning Conference location:Spa, Belgium date:19-20 May 2008
Multi-target models, which predict multiple target variables simultaneously, may predict some of the targets more accurately, and other targets less accurately than a
single-target model. This raises the question
whether it is possible to find, for a given main
target, a subset of the other targets that,
when combined with the main target in a
multi-target model, results in the most accurate model for the main target. We propose Selective Inductive Transfer, an algorithm that automatically finds such a subset.