Skip to main content

007: Identify Automatic Classification cases

Overview​

Identify and document use cases where a file can be automatically labeled based on its content. Examples: documents containing detectable sensitive information using standard Sensitive Information Types, documents matching trainable classifiers related to sensitive topics or document types, etc. Consider different methods for automatic classification with varying assertiveness based on the confidence on the detected data. E.g. for a match to a sensitive information type which is likely to lead to many false positives, use interactive client-side labeling in recommendation mode, whereas for matches with a high level of certainty such as those resulting from Exact Data Matching or fingerprinting at high match thresholds, you can use service-side autolabeling in SharePoint, OneDrive and Exchange. Do not deploy the automatic labeling rules at this stage.

Reference​