Ole Winther - software

ClusterLustre v1.2

Matlab program package for consensus clustering that is combining multiple clustering runs into a single more robust clustering. Described in details in

Robust Multi-Scale Clustering of DNA Microarray Data with the Consensus Algorithm
T. Grotkjær, O. Winther, B. Regenberg, J. Nielsen and L. K. Hansen
Bioinformatics 22, 58-67 (2006)
Download pdf More info including bibTEX + supplementary material ClusterLustre v.1.2 Matlab software

The DTU ICA toolbox

Matlab program packages for various ICA algorithms including the Bayesian mean field algorithm. Described in detail on the The DTU ICA toolbox homepage.

Mean field kernel learning package

A C-program package for mean field kernel learning. Includes ADATAP, naive mean field theory Gaussian process and Support Vector Machines algorithms. Additional features are build-in leave-one-out error estimators and hyperparameter optimization. The hyperparameter optimization schemes, based upon either the mean field estimate of likelihood or leave-one-out error, are not implemented for all algorithms.

References:

Gaussian Processes for Classification: Mean Field Algorithms
M. Opper and O. Winther
Neural Computation, 12, 2655-2684 (2000).
Download ps.Z Download pdf
Efficient Approaches to Gaussian Process Classification
L. Csato, E. Fokoue, M. Opper, B. Schottky and O. Winther
In Advances in Neural Information Processing Systems 12 (NIPS'99), Eds. S. A. Solla , T. K. Leen, and K.-R. Muller, 251-257, MIT Press (2000).
Download ps.Z Download pdf
Gaussian Processes and SVM: Mean Field Results and Leave-One-Out
M. Opper and O. Winther
In Advances in Large Margin Classifiers, Eds. A. J. Smola, P. Bartlett, B. Sch lkopf, D. Schuurmans, 311-326, MIT Press (2000).
Download ps.Z Download pdf

Last updated 19th of Aptil 2002.

Back to Digital Signal Processing homepage

Back to frontpage