Smart Classification Examples

The following examples refer to custom metadata that would have to be created before creating the Smart Classification rule; a Smart Classification rule cannot be saved unless you specify which metadata field to set.

Identifying files less than 5 MB containing US social security numbers 

Rule nameTag files <5 MB with US social security numbers
Automatic executionEnable
FiltersFile size is less than [5 MB]
ClassifierDefault
Classifier patternsMatch pattern by name [U. S. Social Security Number (SSN)]
Classifier conditionNumber of matches is greater than [0]
Match actionSet metadata  [SSN.found] to yes
Non-match actionSet metadata  [SSN.found] to no


Identifying files with extensions .txt and .pdf containing US social security numbers 

Rule nameTag txt and pdf files with US social security numbers
Automatic executionEnable
FiltersFile extension is equal to [txt] OR File extension is equal to [pdf]
Note: Click AND to change it to OR.
ClassifierDefault
Classifier patternsMatch pattern by name [U. S. Social Security Number (SSN)]
Classifier conditionNumber of matches is greater than [0]
Match actionSet metadata [SSN.found] to yes
Non-match actionSet metadata [SSN.found] to no


Identifying all files containing US social security numbers 

Rule nameTag all files with US social security numbers
Automatic executionEnable
FiltersAnything
ClassifierDefault
Classifier patternsMatch pattern by name [U. S. Social Security Number (SSN)]
Classifier conditionNumber of matches is greater than [0]
Match actionSet metadata [SSN.found] to yes
Non-match actionSet metadata [SSN.found] to no


Identifying files in the  Team Folder HumanResources containing US social security numbers 

Rule nameTag all Human Resources files containing US social security numbers
Automatic executionEnable
FiltersFile path starts with [TeamFolderAdmin/HumanResources]

Note: See Identifying a FileCloud Specific Path for help writing FileCloud folder paths.
ClassifierDefault
Classifier patternsMatch pattern by name [U. S. Social Security Number (SSN)]
Classifier conditionNumber of matches [is greater than 0]
Match actionSet metadata [SSN.found] to yes
Non-match actionSet metadata [SSN.found] to no

Identifying files containing any pattern in the custom pattern group France ID numbers

Rule nameFrance ID numbers
Automatic executionEnable
FiltersAnything
ClassifierDefault
Classifier patternsMatch pattern by group [France ID numbers]
Classifier conditionNumber of matches [is greater than 0]
Match actionSet [metadata ID.found] to yes
Non-match actionSet [metadata ID.found] to no

Identifying files with Singapore National Registry Identity Card (NRIC)

Rule nameTag files with Singapore NRIC
Automatic executionEnable
FiltersAnything
ClassifierSingapore NRIC
Classifier conditionNumber of matches [is greater than 0]
Match actionSet [metadata ID.found] to yes
Non-match actionSet [metadata ID.found] to no


Identifying files with patterns matching American Express credit cards

Rule nameTag files with American Express card numbers
Automatic executionEnable
FiltersAnything
ClassifierDefault
Classifier patternsMatch RegEx [3[47]{1}[0-9]{13}]
OR
Match RegEx [3[47]{1}[0-9]{2}-[0-9]{6}-[0-9]{5}]
OR
Match RegEx [3[47]{1}[0-9]{2} [0-9]{4} [0-9]{4} [0-9]{3}]

Note: Smart Classification automatically inserts OR when you add multiple Classifier patterns.
Classifier conditionNumber of matches [is greater than 0]
Match actionSet [metadata ID.found] to yes
Non-match actionSet [metadata ID.found] to no


Identifying files with the exact phrase "Confidential - for internal use only"

Rule nameTag files marked as confidential
Automatic executionEnable
FiltersAnything
ClassifierDefault
Classifier patternsMatch RegEx [Confidential - for internal use only]
Classifier conditionNumber of matches [is greater than 0]
Match actionSet [IncludesText.Confidential] to yes
Non-match actionSet [IncludesText.Confidential] to no


Mark files with different tags depending on the number of matches

In this rule, if a file has 0-2 five-digit numbers, it is marked as having a low possibility of personal ID information. If it has >2 five-digit numbers, it is marked as having a high possibility of personal ID information. This enables you to perform different operations on files with low and high likelihood of having a match. For example, you might choose to manually review files with low possibility, but automatically block files with high possibility.

Rule nameTag files based on number of 5-digit numbers
Automatic executionEnable
FiltersAnything
ClassifierDefault
Classifier patternsMatch RegEx [[0-9]{5}]
Classifier conditionNumber of matches [is greater than 2]
Match actionSet [ID.found] to high
Non-match actionSet [ID.found] to low


Identifying files with a phrase that is the same or similar to "Confidential - for internal use only"

Rule nameTag files with confidentiality phrases
Automatic executionEnable
FiltersAnything
ClassifierSolr Standard Query
Classifier patternsMatch Standard Query ["Confidential - for internal use only"~4]
(include "" around phrase)

Note
: ~4 indicates that all words in the phrase must appear, but may be within 4 words of each other. For example "Confidential - use for internal only" would be a match.
Classifier conditionNumber of matches [is greater than 0]
Match actionSet [IncludesText.Confidential] to yes
Non-match actionSet [IncludesText.Confidential] to no


Identifying files with a word that matches or is one letter different from "Confidential"

Rule nameTag files with words spelled similarly to confidential
Automatic executionEnable
FiltersAnything
ClassifierSolr Standard Query
Classifier patternsMatch Standard Query [Confidential~1]
(do not include "" around word)

Note
: ~1 indicates that there may be 1 letter different in the spelling, for example "Confidental" and "Confidentials" would match, but "Confidentail" would not.
Classifier conditionNumber of matches [is greater than 0]
Match actionSet [Spelling.similar] to yes
Non-match actionSet [Spelling.similar] to no


Identifying files with the word "classified" and not the word "declassified"

Rule nameTag classified files
Automatic executionEnable
FiltersAnything
ClassifierSolr Standard Query
Classifier patternsMatch Standard Query ["CLASSIFIED" NOT "DECLASSIFIED"]
Classifier conditionNumber of matches [is greater than 0]
Match actionSet [Classified.found] to yes
Non-match actionSet [Classified.found] to no


Identifying files marked for blocking by ICAP-DLP

In the case of the ICAP-DLP classifier, the pattern is checked by ICAP-DLP, which tags the file if it is sensitive and does not tag it if it is not sensitive. Therefore, if the file is tagged by ICAP-DLP as sensitive, it is a match, and the following rule sets File.allowed to false, indicating that the file is not allowed to be downloaded, uploaded, or shared.

Rule nameIdentifying files flagged by ICAP-DLP
Automatic executionEnable
FiltersAnything
ClassifierICAP-DLP
Classifier conditionNumber of matches [is greater than 0]
Match actionSet [File.allowed] to false
Non-match actionSet [File.allowed] to true


Identifying files with the names or addresses (AI Classifier example)

Rule nameTag files with names or addresses
Automatic executionEnable
FiltersAnything
ClassifierAI Classifier
Classifier patternsMatch instances of [people names]
OR
Match instances of [addresses]
Classifier conditionNumber of matches [is greater than 0]
Match actionSet [Personal Info.found] to yes
Non-match actionSet [Personal Info.found] to no


Identifying files with company names (AI Classifier example)

Rule nameIdentify files with company names
Automatic executionEnable
FiltersAnything
ClassifierAI Classifier
Classifier patternsMatch instances of [company names]
Classifier conditionNumber of matches [is greater than 0]
Match actionSet [CompanyName.detected] to yes
Non-match actionSet [CompanyName.detected] to no


Identifying files with contact information (AI Classifier example)

Rule nameIdentify files with contact information
Automatic executionEnable
FiltersAnything
ClassifierAI Classifier
Classifier patternsMatch instances of [phone numbers]
OR
Match instances of [email addresses]
Classifier conditionNumber of matches [is greater than 0]
Match actionSet [ContactInfo.detected] to yes
Non-match actionSet [ContactInfo.detected] to no