| SOM Toolbox | Online documentation | http://www.cis.hut.fi/projects/somtoolbox/ |
% SOM Toolbox
Version 2.0beta, Jun 7 2000
Copyright 1997-2000 by
Esa Alhoniemi, Johan Himberg, Juha Parhankangas and Juha Vesanto
Contributed files may contain copyrights of their own.
SOM Toolbox comes with ABSOLUTELY NO WARRANTY; for details
see License.txt in the program package. This is free software,
and you are welcome to redistribute it under certain conditions;
see License.txt for details.
Demos
som_demo1 SOM Toolbox demo 1: basic properties
som_demo2 SOM Toolbox demo 2: basic usage
som_demo3 SOM Toolbox demo 3: visualization
som_demo4 SOM Toolbox demo 4: data analysis
Creation of structs
som_set create & set (& check) values to structs
som_info print out information on a given struct
som_data_struct create & initialize a data struct
som_map_struct create & initialize a map struct
som_topol_struct create & initialize a topology struct
som_train_struct create & initialize a train struct
Struct conversion and file I/O
som_vs1to2 converts a version 1.0 struct to version 2.0 struct
som_vs2to1 converts a version 2.0 struct to version 1.0 struct
som_read_data reads a (SOM_PAK format) ASCII data file
som_write_data writes a SOM_PAK format codebook file
som_write_cod writes a SOM_PAK format data file
som_read_cod reads a SOM_PAK format codebook file
Data preprocessing
som_normalize normalize data set
som_denormalize denormalize data set
som_norm_variable (de)normalize one variable
preprocess preprocessing GUI
Initialization and training functions
som_make create, initialize and train a SOM
som_randinit random initialization algorithm
som_lininit linear initialization algorithm
som_seqtrain sequential training algorithm
som_batchtrain batch training algorithm
som_gui SOM initialization and training GUI
som_prototrain a simple version of sequential training: easy to modify
Clustering algorithms
kmeans k-means algorithm
kmeans_clusters try and evaluate several k-means clusterings
neural_gas neural gas vector quantization algorithm
Supervised/classification algorithms
som_supervised supervised SOM algorithm
lvq1 LVQ1 algorithm
lvq3 LVQ3 algorithm
knn k-NN classification algorithm
SOM error measures
som_quality quantization and topographic error of SOM
som_distortion SOM distortion measure
Auxiliary functions
som_bmus calculates BMUs for given data vectors
som_divide extract subsets of data based on map
som_label give labels to map units
som_autolabel automatically labels the SOM based on given data
som_unit_coords calculates coordinates in output space for map units
som_unit_dists distances in output space between map units
som_unit_neighs units in 1-neighborhood for each map unit
som_neighborhood calculates neighborhood matrix for the given map
som_select GUI for manual selection of map units
som_estimate_gmm create Gaussian mixture model on top of SOM
som_probability_gmm evaluate Gaussian mixture model
som_ind2sub from linear index to subscript index
som_sub2ind from subscript index to linear index
nanstats mean, std and median which ignore NaNs
Using SOM_PAK from Matlab
som_sompaktrain uses SOM_PAK to train a map
sompak_gui GUI for using SOM_PAK from Matlab
sompak_init call SOM_PAK's initialization programs from Matlab
sompak_init_gui GUI for using SOM_PAK's initialization from Matlab
sompak_rb_control an auxiliary function for sompak_*_gui functions.
sompak_sammon call SOM_PAK's Sammon program from Matlab
sompak_sammon_gui GUI for using SOM_PAK's Sammon program from Matlab
sompak_train call SOM_PAK's training program from Matlab
sompak_train_gui GUI for using SOM_PAK's training program from Matlab
Visualization
som_show basic visualization
som_show_add add labels, hits and trajectories
som_show_clear remove extra markers
som_recolorbar refresh/reconfigure colorbars
som_grid visualization of SOM network
som_cplane component planes and U-matrices
som_barplane bar chart visualization of map
som_pieplane pie chart visualization of map
som_plotplane plot chart visualization of map
som_trajectory launches a GUI for presenting comet-trajectories
Auxiliary functions for visualization
som_hits calculates the response of data on the map
som_umat calculates the U-matrix
cca curvilinear component analysis projection algorithm
pcaproj principal component projection algorithm
sammon Sammon's mapping projection algorithm
som_connection connection matrix for map
som_vis_coords map unit coordinates used in visualizations
som_colorcode create color coding for map/2D data
som_bmucolor colors of the BMUs from a given map color code
som_normcolor simulate indexed colormap
som_clustercolor color coding which depends on clustering structure
som_kmeanscolor color coding according to k-means clustering
Low level routines used by visualization functions
vis_patch defines hexagonal and rectangular patches
vis_som_show_data returns UserData and subplot handles stored by som_show.m
vis_valuetype used for type checks
vis_footnote adds a movable text to the current figure
vis_trajgui the actual GUI started by som_trajectory.m
vis_PlaneAxisProperties set axis properties in visualization functions
vis_footnoteButtonDownFcn callback function for vis_footnote.m
vis_planeGetArgs converts topol struct to lattice, msize argument pair
Other
somtoolbox this file
iris.data IRIS data set (used in demos)
License.txt GNU General Public License
Copyright.txt Copyright notice