A way to import an externally trained model into Splunk?

[fatemabwudel]

Hi,

I have a data model trained outside of Splunk using the K-means algorithm and sampled data-set.

I haven't tried the built-in algorithms that come with the MLTK app in Splunk yet, as I am more familiar with using python scripts on my local computer for ML.

Is there a way to import a trained model into Splunk which can then be used to "apply" on new data?

We have Splunk Enterprise Version: 7.3.3.
Any help would be appreciated.

[efavreau]

Maybe someone else knows a way. FWIW, Python is a citizen in the Splunk ecosystem. And MLTK is built on top of the Python for Scientific Computing Library (https://splunkbase.splunk.com/apps/#/search/python%20for%20scientific%20computing/). I understand not wanting to do work again. But the investment to building it in Splunk might be worth considering.

[fatemabwudel]

Thanks efavreau for the information, I saw similar question posted three years ago, and thought that maybe Splunk have that implemented in the newer releases and that it might now support importing already trained models. But looks like it still doesn't support it. I will try to create a model and then train it in Splunk next.