Question & Answer
Question
After I ingested metric data to AIOps MAD, I would like to use AI model management->Metric anomaly detection to train Metric anomaly model. However, in "Set up training" page, I am not able to schedule training on past dates to match the timestamp of my metric data. How do I configure to select metric data that I already ingested for training?
Answer
When you schedule a training and select a date in the calendar, you are not selecting the date of your metric data, but it is the time when you want to run your training. You can deactivate the schedule training and initially run the manual training, which trains all the data that you already ingested to MAD. After the training completed successfully, when new metric data flow to the system, you have to wait for at least 2 weeks or 50% of the latest data available (for example: 7 days, if zero gap) before you can enable daily retraining (schedule training).
[{"Type":"MASTER","Line of Business":{"code":"LOB45","label":"Automation"},"Business Unit":{"code":"BU059","label":"IBM Software w\/o TPS"},"Product":{"code":"SSE9G0Q","label":"IBM Cloud Pak for AIOps"},"ARM Category":[{"code":"a8m3p000000PCG3AAO","label":"Watson AIOps-\u003EAI Manager-\u003EAI Modelling\/Algorithms-\u003ETraining"}],"ARM Case Number":"TS014199361","Platform":[{"code":"PF025","label":"Platform Independent"}],"Version":"All Versions"}]
Product Synonym
AIOps
Was this topic helpful?
Document Information
Modified date:
11 October 2023
UID
ibm17049035