Sample command lines:

1. For running with WRAcc, z-score, kl, and hd measures:

java Test -FILEIN <full path to input file> -FILEOUT <full path to input file> -NUMERIC <number of numeric attributes> -ORDINAL <number of ordinal attributes> -CATEGORIC <number of categoric attributes> -NUMTARGET <number of numeric targets> -ORDTARGET <number of ordinal targets> -CATTARGET <number of categoric attributes> -MEASURE <id of measure>

where

-FILEIN: input file in csv format; in the file the columns have to be arranged in the order: numeric attributes, ordinal attributes, categoric attributes, numeric targets, ordinal targets, categoric targets

-FILEOUT: output file containing the top 50 subgroups together with their scores

-MEASURE: 5 for WRAcc, 6 for z-score, 2 for kl, and 4 for hd


2. For running with qr:

First, please run

java ComputeQRScore -FILEIN <as above> -FILEOUT <output file containing the e term of all scores, assume to be score.txt> -NUMERIC <as above> -ORDINAL <as above> -CATEGORIC <as above> -NUMTARGET <as above> -ORDTARGET <as above> -CATTARGET <as above>

Then, if no sampling is used, please run

java Test -FILEIN <full path to input file> -FILEOUT <full path to input file> -FILESCORE score.txt -NUMERIC <number of numeric attributes> -ORDINAL <number of ordinal attributes> -CATEGORIC <number of categoric attributes> -NUMTARGET <number of numeric targets> -ORDTARGET <number of ordinal targets> -CATTARGET <number of categoric attributes> -MEASURE 1 -EPSILON 0.1

If sampling is used, please run

java Test -FILEIN <full path to input file> -FILEOUT <full path to input file> -FILESCORE score.txt -NUMERIC <number of numeric attributes> -ORDINAL <number of ordinal attributes> -CATEGORIC <number of categoric attributes> -NUMTARGET <number of numeric targets> -ORDTARGET <number of ordinal targets> -CATTARGET <number of categoric attributes> -MEASURE 1 -SAMPLING true -EPSILON <value ranging from 0 to 1>

