[Back]


Talks and Poster Presentations (with Proceedings-Entry):

M. Reiter, T. Melzer:
"Ridge Penalty Regularization for kernel-CCA";
Talk: Austrian Association for Pattern Recognition (ÖAGM), Hagenberg; 2004-06-17 - 2004-06-18; in: "Proceedings of 28th Workshop", Österreichische Computer Gesellschaft, (2004), 33 - 38.



English abstract:
CCA and kernel-CCA are powerful statistical tools that have been successfully employed for feature extraction. However, when working in high-dimensional feature spaces, care has to be taken to avoid overfitting. This paper discusses the influence of ridge penalty regularization on kernel-CCA by relating it to multivariate linear regression (MLR) and partial least squares (PLS).
Experimental results for a pose estimation task are given.


Online library catalogue of the TU Vienna:
http://aleph.ub.tuwien.ac.at/F?base=tuw01&func=find-c&ccl_term=AC04968214

Electronic version of the publication:
http://publik.tuwien.ac.at/files/PubDat_119607.pdf


Created from the Publication Database of the Vienna University of Technology.