M. Yamada, G. Niu, J. Takagi & M. Sugiyama;
JMLR W&CP 20:247–262, 2011.
Computationally Eﬃcient Suﬃcient Dimension Reduction
via Squared-Loss Mutual Information
The purpose of suﬃcient dimension reduction (SDR) is to ﬁnd a low-dimensional
expression of input features that is suﬃcient for predicting output values. In this paper, we
propose a novel distribution-free
SDR method called suﬃcient component analysis
is computationally more eﬃcient than existing methods. In our method, a solution is computed by
iteratively performing dependence estimation and maximization: Dependence estimation is
analytically carried out by recently-proposed least-squares mutual information
dependence maximization is also analytically carried out by utilizing the Epanechnikov kernel
Through large-scale experiments on real-world image classiﬁcation and audio tagging problems,
the proposed method is shown to compare favorably with existing dimension reduction
Page last modified on Sun Nov 6 15:43:46 2011.