Classification by themes is very different compared to both content type and file extension. With themes, a classification is made using a list of predefined keywords that exist within the actual content of the object being stored. Using themes, Amazon Macie can assign more than one theme if multiple themes are detected within the object:
Theme title identifies the type of keywords that are associated, for example, Encrypted Data Keywords. Minimum keyword combinations shows how many words must exist within the object from the associated theme title for it to be classified with the related risk. So, for example, there must be two keyword combinations from within the Big Data Frameworks theme title for it to be associated with that theme and associated a Risk level of 4. If you select Theme title, you can view the keywords that are searched for that theme:
As you can see, the training set keywords that exist for the Big Data Frameworks theme include mapreduce...
