| fldmin | Find the minimum values of the fields |
| -d <srcdata> | source data frame |
| -fout <trg-field> | field to be created |
| [-dout <trgdata>] | target data for statistics in classes |
| [-md <value>] | missing value to be skipped |
| fldmax | Find the maximum values of the fields |
| -d <srcdata> | source data frame |
| -fout <trg-field> | field to be created |
| [-dout <trgdata>] | target data for statistics in classes |
| [-md <value>] | missing value to be skipped |
| fldavg | Compute the average values of the fields |
| -d <srcdata> | source data frame |
| -fout <trg-field> | field to be created |
| [-dout <trgdata>] | target data for statistics in classes |
| [-md <value>] | missing value to be skipped |
| fldsum | Compute the sum values of the fields |
| -d <srcdata> | source data frame |
| -fout <trg-field> | field to be created |
| [-dout <trgdata>] | target data for statistics in classes |
| [-md <value>] | missing value to be skipped |
| fldvar | Compute the variances of the fields |
| -d <srcdata> | source data frame |
| -fout <trg-field> | field to be created |
| [-dout <trgdata>] | target data for statistics in classes |
| [-md <value>] | missing value to be skipped |
| flddev | Compute the standard deviations of the fields |
| -d <srcdata> | source data frame |
| -fout <trg-field> | field to be created |
| [-dout <trgdata>] | target data for statistics in classes |
| [-md <value>] | missing value to be skipped |
The field statistics compute statistical values from fields. These commands reduce the information of each field to one quantity, which is stored in a given field. If the output data frame has been specified, then the operation creates a data frame (if it does not exist) and adds the field into it. Otherwise, the new field is created directly into specified (or current) directory.
The form of the result is a column matrix in which one statistical value over all the fields is stored in one field. Each data record corresponds to a field in the source data. This differs from the class statistics in which data records correspond to original classes.
Each operation may take a missing value as a parameter. If the parameter is specified, then the defined value is skipped while computing the statistics.
Example (ex4.7): A typical use of field statistics is to compute minimum and maximum values or average and deviation from a frame for some scaling task.
... NDA> fldstat -d boston -dout mmstat -min -max NDA> ls -fr mmstat mmstat.min mmstat.max NDA> fldstat -d boston -dout adstat -avg -dev NDA> ls -fr adstat adstat.avg adstat.dev