When producing a clustering model it seems that Augustus will modify the
contents of the Array element from the input PMML but will not check to make
sure that the "n" attribute is correct.
In my case, the input PMML looked like the first snippet, but I had thirty
clustering fields. The output PMML looked like the second snipppet.
--input--
<Cluster>
<Array type="real" n="2">1 2</Array>
</Cluster>
--output--
<Cluster id="12" size="43">
<Array type="real" n="2">0.0543841445374 0.0485577351248 0.0442172816649 0.0268333242814 0.0258044444262 0.411797453267 0.175113935481 0.0786815710599 0.0348041127382 0.02136878921 0.518328703764 0.490628080134 0.476526913056 0.437696497242 0.30651546755 0.0440096238174 5.96689134715e-05 5.12650252702e-05 2.90417020953e-05 6.76034843594e-05 0.453176119483 0.152045421402 0.0998874300404 0.0645827476412 0.0498775338892 0.0 0.0232558139535 0.0232558139535 0.0232558139535 0.0232558139535</Array>
</Cluster>
Original issue reported on code.google.com by dlo...@opendatagroup.com on 13 Apr 2012 at 4:46
Original issue reported on code.google.com by
dlo...@opendatagroup.com
on 13 Apr 2012 at 4:46