Matlab routines



This is a list of the main MATLAB routines you can use to develop Kohonen and CP-ANNs models

model_gui

model_gui opens a GUI for calculating both Kohonen Maps and CPANNs; in order to open the graphical interface, just type on the matlab command line:

model_gui

there are no inputs, data can be loaded and saved directly from the graphical interface.

[-> top]


som_settings

default setting structure for Kohonen maps and counterpropagation artificial neural networks (CP-ANNs); som_settings build a default structure with all the parameter needed to perform Kohonen maps, CP-ANNs, XYF and SKN models

settings = som_settings(type);

type "help som_settings" on the MATLAB command window for further informations

[-> top]


opt_model

opt_model searches the optimal number of neurons and epochs for classification model based on SOMSs by means of Genetich Algorithms

opt_res =
opt_model(X,class,settings,val_type,num_groups,opt_fun,ns_bank,ep_bank,rep_model);


type "help opt_model" on the MATLAB command window for further informations

[-> top]

model_kohonen

Kohonen maps; model_kohonen builds Kohonen maps (self organising maps, SOM)

model = model_kohonen(X,settings);

type "help model_kohonen" on the MATLAB command window for further informations

[-> top]


model_cpann

Counterpropagation Artificial Neural Networks (CP-ANNs); model_cpann builds a classification model based on CP-ANNs

model = model_cpann(X,class,settings);


type "help model_cpann" on the MATLAB command window for further informations

[-> top]


model_skn

Supervised Kohonen networks (SKN); model_skn builds a classification model based on Supervised Kohonen networks

model = model_skn(X,class,settings);


type "help model_skn" on the MATLAB command window for further informations

[-> top]


model_xyf

XY-fused networks (XY-F); model_skn builds a classification model based on XY-fused networks

model = model_xyf(X,class,settings);


type "help model_xyf" on the MATLAB command window for further informations

[-> top]


pred_kohonen

prediction of unknown samples with Kohonen maps; pred_kohonen projects new samples by using a previuos model built by means of Kohonen maps (model_kohonen)

pred = pred_kohonen(X,model);

type "help pred_kohonen" on the MATLAB command window for further informations

[-> top]


pred_cpann

prediction of unknown samples with counterpropagation artificial neural networks (CP-ANNs); pred_cpann predicts classes of unknown samples by using a previuos model built by means of CP-ANNs (model_cpann)

pred = pred_cpann(X,model);

type "help pred_cpann" on the MATLAB command window for further informations

[-> top]


pred_skn

prediction of unknown samples with supervised Kohonen networks (SKNs); pred_skn predicts classes of unknown samples by using a previuos model built by means of SKNs (model_skn)

pred = pred_skn(X,model);

type "help pred_cpann" on the MATLAB command window for further informations

[-> top]


pred_xyf

prediction of unknown samples with XY-Fused networks (XY-Fs); pred_xyf predicts classes of unknown samples by using a previuos model built by means of XY-Fs (model_xyf)

pred = pred_xyf(X,model);

type "help pred_cpann" on the MATLAB command window for further informations

[-> top]


cv_cpann

cross validation for Counterpropagation Artificial Neural Networks (CP-ANNs); cross validation is performed with venetian blinds (i.e. with 3 cv groups the split of the first group will be [1,0,0,1,0,0,....,1,0,0] and so on) or contiguous blocks (i.e. the split of the first group will be [1,1,1,1,0,0,....,0,0,0] and so on).

cv = cv_cpann(X,class,settings,cv_type,cv_groups);

type "help cv_cpann" on the MATLAB command window for further informations

[-> top]


cv_skn

cross validation for Supervised Kohonen networks; cross validation is performed with venetian blinds (i.e. with 3 cv groups the split of the first group will be [1,0,0,1,0,0,....,1,0,0] and so on) or contiguous blocks (i.e. the split of the first group will be [1,1,1,1,0,0,....,0,0,0] and so on).

cv = cv_skn(X,class,settings,cv_type,cv_groups);

type "help cv_skn" on the MATLAB command window for further informations

[-> top]


cv_xyf

cross validation for XY-fused networks; cross validation is performed with venetian blinds (i.e. with 3 cv groups the split of the first group will be [1,0,0,1,0,0,....,1,0,0] and so on) or contiguous blocks (i.e. the split of the first group will be [1,1,1,1,0,0,....,0,0,0] and so on).

cv = cv_xyf(X,class,settings,cv_type,cv_groups);

type "help cv_xyf" on the MATLAB command window for further informations

[-> top]


visualize_model

visualisation of the Kohonen map or CP-ANNs results; visualize_model opens a GUI figure for exploring the results

visualize_model(model);

Sample labels and variable labels can be loaded as optional inputs. Type "help visualize_model" on the MATLAB command window for further informations

[-> top]