By using General Linear task in statistica, Can we keep constant beta values in PMML script

[Manoj Katasani]

Hi All,

I have created a model usingGeneral Linear regression task in TIBCO statistica. Every run a new set of beta values are generated in PMML Script .

Can we keep constant Beta values in PMML script..

Note :- I can get this done by, without usingGeneral Linear regression task , by using contant PMML task and fixed values. But i need by usingGeneral Linear regression task. Please let me know your thoughts on this.

Many thanks for your time. happy learning .

Regards

Manoj.K

[Tomas Jurczyk]

Hi.

Setting up PMML in the way as it is in the shared workflow screanshot means that PMML is always created after each run again. Modelling node called "General Linear" will do that (it is modelling node, which is doing modelling each run, modelling is in our case preparation of PMML with parameters).

Solution to that is to use workspace similar to that in attachment. PMML node can have code stored, it can take it from savedPMML on thedisk or from the Enterprise meta-repository or simply have only code written in the node and not saved anywhere. So, I believe solution would be to save PMML to disk or Enterprise to reference it or simply delete "General Linear" node to have workflow to score new data and not to create a new model.

I hope this helps.

Tomas

[Manoj Katasani]

Great thanksTomas it really helps a lot.

Regards

Manoj.K

[Manoj Katasani]

Great thanks Tomas it really helps a lot.

 

Regards

 

Manoj.K