Tuning a sensitive information type's confidence level
When you're creating a DLP policy, you can use sensitive information types (such as
Project Wildebeest in the previous recipe). Matches are based on specific patterns and confidence level. The higher the confidence level, the more certain the match must be whereas a lower confidence level will catch more similar matches (though these will include false positives).
You may wish for a higher confidence level when you're more absolutely certain data that should be caught follows a specific pattern each time, such as a routing number within a certain proximity of an account number or SSN. Lower confidence levels will be more inclusive of potential matches and may be helpful for more loosely structured data where, perhaps, a confidential project may be discussed or referenced in numerous ways and structures.
In this recipe, we'll adjust a sensitive information type from the previous recipe (