How automatic acceptance and rejection works
During the data classification process, Collibra predicts the data class for a column. The percentage next to the data class indicates the confidence level of the data classification suggestion.
As a data steward, you can manually accept or reject a suggested data class, but you can also configure thresholds to automatically accept or reject the data classification suggestions. The automatic acceptance or rejection reduces manual review efforts.
- If you set the automatic acceptance threshold to 75%, then a data classification suggestion with a confidence level of 75% or higher is accepted automatically.
- If you set the automatic rejection threshold to 49%, then a data classification suggestion with a confidence level of 49% or lower is automatically rejected and does not appear for the column.
Important considerations
-
Once a data classification has been accepted automatically for a column, the data classification won't be automatically updated if you run the data classification process again.
-
If the acceptance threshold and rejection threshold are set to the same value, and a data classification suggestion has this confidence level percentage, the classification suggestion will be rejected.
-
If multiple classification suggestions meet the threshold for a column, the suggestion with the highest confidence level percentage is accepted automatically, as long as this suggestion is the only one to have that confidence level percentage. If two or more suggestions have the same confidence level, none are accepted automatically, and all remain visible.
ExampleYou set the automatic acceptance threshold to 85% and classify a table with 2 columns.
- For column A, there are 3 classification suggestions with confidence level 93%,92%, and 90%.
- For column B, there are 2 classification suggestions with the same confidence level of 86%.
The result of the automatic acceptance is:
- For column A, the classification suggestion with 93% is accepted automatically.
- For column B, both suggestions remain visible, none are accepted automatically.
Tip Start by manually accepting and rejecting a suggested data class. Only switch to automatic acceptance and rejection if you are comfortable with the data classification results.
Configure thresholds to automatically accept or reject