Introduction

This module describes how to optimize ML models used by BigFix Runbook AI components like iRecommend and iUnique for recommending relevant runbooks and ticket clustering. It explains how to configure hyperparameters and conduct experimentation to achieve an optimized model.

This module covers the procedure for optimizing the machine learning-based models used by BigFix Runbook AI components like iRecommend and iUnique for recommendation of relevant runbooks and ticket clustering, respectively. This module becomes helpful when organizations feel that the accuracy of recommendations needs further improvement. The issues could be related to model hyperparameters where configuration manager might not have configured correct values before using the same in a production environment.

BigFix Runbook AI provides configurational capabilities where user can define or select a combination of algorithms and their parameter values. These values known as hyperparameter templates are used to check their applicability in particular customer environment via Workbench Analysis. In workbench analysis, user can upload a sample set of ticket descriptions to system and verify whether configured hyperparameters for iRecommend and iUnique is providing accurate results. If results are not as expected, experimentation can be done by changing the parameter values to arrive at an optimized model.

Let’s begin with the configuration of hyperparameters for iRecommend and iUnique.