Interactive method of using a group similarity measure for providing a decision on which groups to combine

Claim 1

A computer-implemented method of using a group similarity measure, for interaction with an analyst, on a plurality of groupie, wherein each of said groups contains software features that can be compared to software features in other groups, said method comprising the following steps: (a) computing a similarity value between pairs of said groups, using said similarity measure to identify at least one similar pair of groups; (b) storing said similar pair of groups as computed by said similarity measure; and advancing to step (d) if the similarity value is zero; (c) determining whether said similar pair can be accepted, as follows: if the similarity value for said pair is not zero, set the pair aside and repeat step (c); c1) comparing said pair to a predetermined set of pairs identified as a never-list, if a never-list pair on the never-list is embedded in said pair being compared, that is, said pair being compared is a superset of a corresponding group in the never-list pair, set said pair aside, and repeat step (c); c2) comparing said pair to a predetermined set of groups identified as a suspension list, if a group in said pair is on the suspension list, set this pair aside, and repeat step (c); c3) determining whether the two groups associated with said pair have different labels, wherein said labels are compared to a listing of labels on a same-label list, whereby if one of the groups associated with said pair is on the same-label list, set this pair aside and repeat step (c); c4) if (none of the rules above) steps c1, c2 and c3 do not apply, and said pair has a combined group size less than or equal to a predetermined minimum group size: merge the two groups associated with said pair into a single new group; c5) label the new group with a label previously associated with the larger of the two groups associated with said pair; go back to step (a); c6) if said new group is greater than a predetermined maximum automatic group size, skip to step (d); c7) if said new group is greater than a predetermined maximum automatic group size, set this pair aside, and repeat step (c); (d) display said pair of groups on a screen and display a group having a similarity value closest to said pair; and also display said pair, a group having label that is most similar to each of the two grgups associated with said pair.