Contact...


Feel free to write...!
You can either login to see email address - or use form below...












Report bug...


Before sending a bug report please consider the following...:

  • If your decision tree is utilizing surrogate rules in case of null input values, then that may cause SQL scores that are different from your training scores. The pmml2sql parser does not yet support surrogate rules, but will instead retrieve scores from default (previous) node.
  • There is a known issue regarding neural nets in Rattle. The Z values in the sigm(Z) function sometimes get out of bounds and causes invalid scores. The basic structure of the Rattel neural net PMML is however not different from other similar neural nets (that are working), so until evidence to the contrary, this is regarded as an error in the PMML/model.
  • Is your model among the ones that are currently integrated into the parser? If not, then feel free to request it...
  • If you have in any way altered the input variables within the data mining tool - and not made similar alterations in your score data - then the input variables will differ between your training data in the datamining tool and the score data in your database. And your final scores will differ as a consequence. For that reason it is a good idea to do all variable manipulations in the database environment, and not in the data mining tool.

Please log in first - to get email address for bug report. Thanks!


Request new model...


Try to upload the PMML first! It might just work....

Please log in first - to get email address. Thanks!