Enhancements to Propose Classification Code Action
This feature delivers several enhancements to the Propose Classification Code action within embedded machine learning:
- Simplifies the Propose Classification Code action
- Validates that proposed classification codes are active prior to saving
- Excludes proposed classification results with 0% probability from display
- Identifies assigned classification codes
Simplifies the Propose Classification Code action
The Propose Classification Code action has been streamlined to enhance the user experience. The Model Type and Domain fields have been removed, simplifying the process. Additionally, a new Threshold% field is now available, allowing users to specify a minimum threshold for displaying results.
Propose Classification Code Action Input
The action has been updated to run synchronously, allowing GTM to run the action in the background and automatically display the results when available. Previously, you needed to click Refresh to view the action results.
Propose Classification Code Action Results
Validates the Proposed Classification Codes are Active
When saving a proposed classification code to an item, GTM now verifies that the code being assigned is active. If a classification code was valid at the time the model was trained, but a new data version is downloaded before the propose classification code action is triggered, the code may become invalid. Before saving the proposed classification code, GTM checks that the code is still valid. If it is not, GTM displays a message indicating that the classification code is no longer valid.
Excludes Proposed Classification Results with 0% Probability from Display
GTM no longer displays proposed classification codes that have a 0% probability of being correct.
Identify Assigned Classification Codes
When the Propose Classification Code action assigns a classification code to an item, the note 'Classification Code proposed by AI' is added in the Item Classification > Product Classification Notes.
Product Classification Notes
Business Benefit: These enhancements to the Propose Classification Code action improve both efficiency and accuracy in classification processes. By streamlining the action, ensuring only active codes are proposed and saved, filtering out results with 0% probability, and clearly identifying already assigned codes, users can make faster, more informed decisions. This reduces errors and rework, enhances compliance, and increases overall productivity for businesses using embedded machine learning for classification tasks.
Steps to Enable and Configure
You don't need to do anything to enable this feature.
Tips And Considerations
When the Propose Classification Code action in embedded machine learning is enhanced, it is recommended to retrain your existing machine learning projects.