We present in this paper a new method for blind source separationwhich is adapted to the case where the sources have different morphologies.We show that the morphology diversity concept leads to anew and very efficient method, even in the presence of noise.




File list:
datafile
.......\iris-solution.txt
.......\iris.txt
.......\leuk72_3k.txt
agg_hierarchical.m
Copyright.txt
CVAP-examples.zip
daisy.m
daisyc.dll
data_normalization.m
find_nearpoint.m
ind2cluster.m
License_Netlab.txt
mainCVAP.fig
mainCVAP.m
mainCVAP6.fig
mainCVAP6.m
neural_gas.m
pam.m
pamc.dll
pca_of_data.m
plot_data_bylabels.m
Readme.txt
show2dim_byclass.m
similarity_euclid.m
similarity_pearson.m
similarity_pearsonC.m
som.m
somtrain.m
som_netlab.m
valid_clear_clustering.m
valid_clear_plotting.m
valid_clear_validation.m
valid_clusteringAlgs.m
valid_data_load.m
valid_data_plot.m
valid_data_plot_check.m
valid_data_split.m
valid_DbDunn.m
valid_errorate.m
valid_errorate_check.m
valid_external.m
valid_findk.m
valid_hierarchical.m
valid_internal.m
valid_internal_deviation.m
valid_internal_intra.m
valid_intrainter.m
valid_mutiple_validation.m
valid_plot_hierarchytree.m
valid_plot_index.m
valid_print_label.m
valid_redraw.m
valid_runclustering.m
valid_runvalidation.m
valid_save_solution.m
valid_solution_load.m
valid_sumpearson.m
valid_sumsqures.m
valid_validation_check.m
xlim2.m