predictive intelligent change capability when train

[Cherly]

After the upgrade to Vancouver. when we train the capability change to from classification to Workflow Classification. is there any way to change fix this issue.

[Brian Bakker]

@sakshi arora 

We have documented the new ML APIs with examples in Using Machine Learning APIs

[sakshi arora]

Hi @Brian Bakker ,

Thanks for the previous response. We have now updated our ML solutions to new ML capability of Workflow classification and have got one issue post we updated : previously training schedules were running automatically based on the training frequency being set. But since we have updated the solutions with workflow classification capability, solutions are not getting re-trained on its own. Do we have to do any other specific change to auto train these solutions in the new capability?

[Brian Bakker]

@sakshi arora 

Workflow Classification solutions should be no different when it comes to auto-training them. However, we did fix an issue when upgrading to Washington that did cause auto-training of solutions to stop working. In this scenario, you will need to click on the "Update & Train" button, and it will now create a new Training Schedule, and auto-retrain the Workflow Classification solution based on the Training Frequency set in the solution definition. We apologise for the inconvenience caused, but auto-training is the same for both Classification and Workflow Classification solutions.

Hope this helps.

Regards,

Brian

[sakshi arora]

Seems re-training manually helped resolved the issue. Thankyou!

[k_lutz]

Have to say that our solution gets the wrong prediction most of the time. Taking a look at the link above but not written well to follow when trying to follow incorrect predictions on almost every test.