%0 Conference Proceedings %T Non-linear feature extraction by the coordination of mixture models %+ Instituut voor Informatica (IvI) %A Verbeek, Jakob %A Vlassis, Nikos %A Krose, Ben %< avec comité de lecture %B 9th Annual Conference of the Advanced School for Computing and Imaging (ASCI '03) %C Heijen, Netherlands %P 287--293 %8 2003-06-04 %D 2003 %K Canonical Correlation Analysis %K Principal Component Analysis %K Self-organizing Maps %Z Computer Science [cs]/Machine Learning [cs.LG]Conference papers %X We present a method for non-linear data projection that offers non-linear versions of Principal Component Analysis and Canonical Correlation Analysis. The data is accessed through a probabilistic mixture model only, therefore any mixture model for any type of data can be plugged in. Gaussian mixtures are one example, but mixtures of Bernoulli's to model discrete data might be used as well. The algorithm minimizes an objective function that exhibits one global optimum that can be found by finding the eigenvectors of some matrix. Experimental results on toy data and real data are provided. %G English %2 https://inria.hal.science/inria-00321490v2/document %2 https://inria.hal.science/inria-00321490v2/file/verbeek03asci.pdf %L inria-00321490 %U https://inria.hal.science/inria-00321490