What are the classification of data?

Data classification approaches — which include categorization by type, owner, regulation, sensitivity, and retention requirement — enable organizations to focus their security, privacy, and analytics efforts on their most important datasets. There are several methods for classifying data, which can be broken into two broad groups: automated and manual: • Automated: Automated approaches rely on some combination of content (keywords, phrases, regular expression searches, etc.), context (who the user is, what their role is, where the file is located, where is it is accessed, what purpose was the data collected for, etc.) and environment (file size, format, owner, creation date, etc.) to make classification decisions. o NOTE: Although many vendors offer automated data classification tools that can classify more data more accurately, while minimizing user effort, they are not 100% accurate, especially if they use machine learning or artificial intelligence algorithms for which models require ongoing training. • Manual: Human classification can be more accurate but is generally reserved for newly created data and is completely reliant on the user’s understanding of classification policies and their ability to apply them properly and consistently. o NOTE: Data classification initiatives have often failed because they were dependent on manual efforts by users with insufficient training.