Boosting with Incomplete Information
Published on Aug 06, 20083564 Views
In real-world machine learning problems, it is very common that part of the input feature vector is incomplete: either not available, missing, or corrupted. In this paper, we present a boosting approa