How to check the accuray of forecast made using machine learning in kibana?

I wanted to know that while doing forecast using log data , is there any way by which i can show how much accurate forecast is so that i can present it in my sprint.
and what all algorithm are used in forecasting

The algorithms to create the forecast models are the same as what is used for anomaly detection. It is too complicated to describe here, but you can get some info by watching this video from ElasticON: https://www.elastic.co/elasticon/conf/2018/sf/the-math-behind-elastic-machine-learning

The accuracy of the forecast is going to depend on the amount of data used to build up the model(s) in the first place. One way to "test" the accuracy is to run a job to build up the model, then run a forecast, then allow the ML job to continue to see the new data that overlaps with the prior forecast. The delta between the two will show you the "accuracy".

yeah, thats a good way to check predicted accuracy of it... I will monitor on my logs forecast and will revert back if any other issues i got. Thank you so much sir , you are really a good person

1 Like

This topic was automatically closed 28 days after the last reply. New replies are no longer allowed.