Splunk IT Service Intelligence

ITSI predictive score question


We have not been able to deploy Predictive and anomaly detection capabilities that can leverage MLTK as current ML algorithms in ITSI cannot account for business hour/seasonality.

The question that I have is... is there a query example that you can point me to that would show how to exclude specific date ranges such as day of the week and hour for example exclude sat and Sunday.

We are trying to include in glass tables predictive score, but the thing is with the current capabilities of ITSI it's not possible to exclude time ranges.  What we want to do is leverage MLTK for that.. to create custom prediction health score model... but with exclusion of time ranges.... then publish that model back into ITSI and show health score..

I guess the question is... how do I exclude time ranges from considerations... I guess the better question is... when ITSI predicts the health score does it take into consideration time data?  If so how can I use the same algorithm, but by excluding time ranges...?  

If you can point me to appropriate resource, this would greatly help me out.

Labels (1)
0 Karma