topic Re: Combine outputs of different platform and use as an input to Ensemble Node in SAS Data Science

Combine outputs of different platform and use as an input to Ensemble Node

kewldude — Wed, 01 Nov 2017 20:51:35 GMT

Hi,

Lets say I used a decision tree and logistic regression nodes in SAS Enterprise Miner and then I have another model produced using SVM in Python . How can I get the python output and use it in SAS EM with the intention of using the output of the decision tree, logistic regression and this SVM from Python as an input to the Ensemble Node?

I know SAS EM has an SVM node and the probable answer to my question is why use Python when you can use SAS EM and its SVM node. Still, the question is if i want to combine outputs from a different platform/language and use it in SAS EM, how do I go about doing that? Any samples?

Re: Combine outputs of different platform and use as an input to Ensemble Node

Reeza — Wed, 01 Nov 2017 21:08:30 GMT

My brain is running a blank right now, but there's a model structure that SAS can import that many languages support, something like PLC/PLM/PML?

EDIT: PLM, R has a package for this as well, not sure about Python.

Re: Combine outputs of different platform and use as an input to Ensemble Node

WendyCzika — Thu, 02 Nov 2017 13:13:29 GMT

There are instructions and a sample SAS Enterprise Miner diagram (the XML file) at this GitHub repo:

https://github.com/sassoftware/enlighten-integration/tree/master/SAS_EM_PythonIntegration

This shows how you can run your Python code from a SAS Code node, then use the Model Import node to be able to assess your model as you can with SAS Enterprise Miner models.