Record Categorizations

To review record categorization assignments, navigate to the Categorizations page of a Categorization project.

To accept many suggested categories at once:

1. Select one or more records.
2. Choose Verify and then choose Tamr Suggestions or Expert Suggestion.

To investigate conflicting responses between Tamr and experts or among experts:

1. Select a single record.
2. On the top right hand side, choose the pop-out tab and then choose Details.
3. Choose a categorization.

Verifying Categorizations

If Tamr and Reviewers have differing opinions, curators can choose which suggestion to accept. To decide which suggestion to accept, use the Verify option for one or more selected records at a time.

• The Top Expert Suggestion accepts the expert categorization with the most upvotes. In case of a tie between experts, the categorization remains unverified.
• The Tamr Suggestion accepts the categorization suggested by the Tamr machine learning model.

Resolving Conflicting Responses

If conflicting categories exist, you as the curator, can choose which one to accept.

In the previous screenshot, the green checkmark indicates the category chosen by the curator. If, as a curator, you wish to switch the category, choose the gray circle next to the category. Once a categorization is verified, Tamr includes it in its machine learning model.

Removing Categorizations

Removing a record categorization

To remove a record user-categorization:

1. Navigate to the Categorizations page of a Classification Project.
2. Select a record.
3. On the top right hand side, choose the pop-out tab, and then choose Details.
4. Choose New Categorization > Clear Categorization.

Clearing all categorizations for a single category

Wehn you remove user categorizations for a selected category, Tamr removes categorizations for all of its descendants.

To remove all record user categorizations for a selected category:

1. Navigate to the Categories page of a Classification Project.
2. Select a category.
3. Choose the three vertical dots icon in the upper right and then choose Clear Categorizations.

Applying Feedback and Updating Categorization Results

The Apply feedback and update results job:

• Trains the model using the latest verified record categorizations.
• Applies the model to all records.
• Calculates record categorization confidences and their average.

To run the Apply feedback and update results job, the following conditions must be met:

• Your project must have one or more unified attributes specified as Included in machine learning.
• You must associate the project with a taxonomy.
• You cannot run any other job at the same time.
• The unified dataset must be up-to-date.

To run the Apply fedback and update results** job:

1. Navigate to the Categorizations tab of a Categorization project.
2. Choose Apply feedback and update results.

Updating Categorization Results

The Update Categorizations job:

• Applies the model to all records.
• Calculates record categorization confidences and their average.

You can run the Update results job only after at least one Apply feedback and update results job has been completed, or after a model has been imported for the project.

To update categorization results:

1. Navigate to the Categorizations tab of a Categorization project.
2. Choose the arrow on Apply feedback and update results, and then select Update results only.