codetext -f <str-fld> -dout <codes> [-split | -prob]
- Code a string field into SOM training data. -split and -prob provide
different coding methods
weight2str -d <codes> -fout <strs-out>
- Convert a default or -prob text training data or the weights of a
trained SOM into strings
instext -dout <dataout> -text <str1> <str2> ... <strn>; [-sen]
- Code a string specified on the command line into a string field
descand -f <str-fld> [-fout <strs-out>]
- Remove Scandinavian alphabet from a field of strings. If <strs-out> is
specified, the data is copied into a new field
textprob -f <str-fld> -dout <prob-matr> [-aout <avg-fld>]
- Calculate conditional probability matrix indicating how characters follow
each other within the words of a string field
text3dprob -f <str-fld> -dout <prob-matr> [-aout <avg-fld>]
- Calculate a 3-D conditional probability matrix indicating how characters
follow each other within the words of a string field. Two previous characters
are taken into account
condprob -f <intfld> -dout <prob-matr> [-c <cldata> -aout<avg-fld>]
- Calculate conditional probability matrix indicating how integers follow
each other within a field. The order of the integers can also be specified
as a classification, in which case the beginning and end of each class is
noted
condent -d <prob-matr> -fout <ent-value>
- Calculate the entropy of a conditional probability matrix
probent -d <prob-matr> -fout <ent-value>
- Calculate the entropy of a simple probability matrix
prob2str -d <prob-matr> -fout <str-value> [-avg <avg-len>]
- Convert a conditional probability matrix into a string of characters
3d2str -d <prob-matr> -fout <str-value> [-avg <avg-len>]
- Convert a 3-D conditional probability matrix into a string of characters
pd2cld -d <prob-matr> -cout <cldata> [-avg <avg-len>]
- Convert any conditional probability matrix into a classification
complexity -d <prob-matr> -fout <compl-value> -acc <accuracy>
- Calculate the amount of bits needed to represent some data using a
specified accuracy
difference -p <prob-matr1> -q <prob-matr2> -fout <diff-value> [-acc <accuracy>]
- Calculate the Kullback-Leibler difference of two probability distributions
modelcomp -s <som> -d <datain> -c <som-cld> -dout <results> [-acc <accuracy> -rout <res-histo> -sout <som-histo> -hout <full-histo> -histo -frac]
- Evaluate the model complexities of a TS-SOM representation of some original
data using a specified accuracy or histogramming. The histograms can be
obtained for later evaluation and the results can be calculated as true
bits or fractions of bits
codesent -c1 <cldata1> -c2 <cldata2> -dout <codes> [-bin -step]
- Code the sentences of some text data using neuron indexes of a word SOM.
Each class of <cldata1> specifies one sentence, and <cldata2> is one
layer of a word map. The result can be binary or gradually descending, which
also indicates word order