Histograms in subgraphs

 

hst	Create a histogram
<graphic>	name of the graphic structure
[-inx <index>]	place of the bar related to others
-d <field>	source field or data frame
[-x <xaxis>]	field for defining the x axis
[-sca <scaling>]	scaling function, default=DATA
[-min <min>]	minimum value
[-max <max>]	maximum value
[-co <color>]	color for the histogram bars
hst <graphic> -d <data> -rm	Remove a histogram

This command sets a histogram for subgraphs. The actual histogram data can be computed with fldhisto and clhisto (see sections 4.5.2 and 4.6.6). These commands produce two different formats for histogram data, which must be taken into account when a graphical presentation is created.

fldhisto

This command creates a frame with two fields: x and N. Thus the command hst should refer to field N. After one histogram has been created, the index must be specified.

clhisto

This command creates a data frame, in which a class (neuron) has one data record and each bin of the histogram is stored in its own field. These fields are named with running numbers from 0 to . In addition, the histogram is a frame containing this data frame called data and a field for the x axis named x. hst does not require parameter -index. All the subgraphs will get a histogram. Of course, this command requires that a graphic structure has a subgraph for each class as SOM based graphics should always have.

Example (ex7.23): This example shows, how histograms can be set for each neuron in a SOM.

# SOM training and classification
...
NDA> clhisto -c c1 -max 10 -f boston.age -dout hst
NDA> mkgrp grp1 -s s1
NDA> layer grp1 -l 3
NDA> hst grp1 -d hst.data -co red -sca som2

Example (ex7.24): This second example shows, how a simple histogram can be computed for one field. For another example, see the description of command fldhisto in section 4.5.2.

# SOM training and classification
...
NDA> fldhisto -f boston.age -max 30 -dout fhst
NDA> mkgrp grp
NDA> addsg grp
NDA> hst grp -inx 0 -d fhst.N -co red