A New Method of Model Selection Based on Learning Coecien t
Keisuke Yamazaki, Kenji Nagata, Sumio Watanabe · 2005
In the information engineering eld, many practical learning machines, e.g. neural net- works, mixture models and hidden Markov models, have been developed. In spite of their wide-range ap- plications, there is no theoretical method to select the optimal sized model in such models i.e. singular mod- els. Recent years, an approach to analyze the singular models was established based on algebraic geometry. In this paper, we propose a new model selection crite- rion, Singular Information Criterion (SingIC), based on the algebraic geometrical method.