Procedure Cluster() does not work as expected.

Hi!

I have some problems with procedure Cluster().

I try simple example but procedure does not work as expected.

I have some stores for CATG "1" with the following values:

    1,100003,4

    1,100006,4

    1,100010,7

    1,100028,7

    1,100058,10

    1,100087,13

Accordind to them I calculate i_meas as each of them divided by average

i_meas:

    1,100003,0.53333333333333

    1,100006,0.53333333333333

    1,100010,0.93333333333333

    1,100028,0.93333333333333

    1,100058,1.3333333333333

    1,100087,1.7333333333333

Then I run the rule:

    PointMembership:out_meas, Centroid:out_cent <-Cluster(Measure:i_meas, Method:"BANG", NumClusters:4, ClusterHier:"LOC", ClusterOverHier:"PROD", ByGroupDims:"CATG")

   

There are i_meas on intersection catg/str_ (which contain 6 rows mentioned above) and argument "NumClusters" that is 4. So I expect that rule returns 4 cluster but out_cent measure contains the folowing: