The requirement is to do forecasting on indexed data. A python script will be developed and use in Splunk to use the indexed data for forecasting. Is this possible? if yes, how?
Thanks!
@teddyidc1101 if you are using scikit-learn, pandas, statsmodel, numpy, or scipy libraries
as forecasting algorithm you should check out Splunk Machine Learning Toolkit (MLTK) and if any of currently supported ML libraries are not already present in MLTK you can extend and import your own algorithm using ML SPL API
Also refer to State Space Forecast algorithm introduced in 4.2 which allows you to fit and apply learnt model for time series forecasting: https://docs.splunk.com/Documentation/MLApp/latest/User/Algorithms#StateSpaceForecast
Yes it is possible you can create custom command and pass indexed data as a parameter