Optimized file classification with supervised learning
A method of classifying a file, including extracting metadata from the file, assigning a classification for the file by applying a machine learning model that was trained to classify files based on the metadata, determining a confidence level representing an accuracy of the classification, wherein if the confidence level is below a threshold value analyze the content of the file to assign a classification for the file based on the content; and store the assigned file classification.
1. A method of classifying a file locally at a computer, comprising:
extracting metadata from the file;
assigning a classification for the file by applying a machine learning model that was trained to classify files based on the metadata; wherein the machine learning model is provided by a remote server to classify files locally at the computer;
determining a confidence level representing an accuracy of the classification;
wherein if the confidence level is below a threshold value analyzing the content of the file to assign a classification for the file based on the content;
storing the assigned file classification; wherein the classification is stored as metadata in the file; and
permitting or preventing access to the file by a user of the computer responsive to the assigned classification;
wherein at least one classification requires a second factor authentication by the user, wherein the user is required to enter a onetime code received in an SMS or Email in addition to entering a password.
2. The method of claim 1 , wherein the metadata is enriched with information related to file creator metadata or file editor metadata, including risk levels assigned to them.
3. The method of claim 1 , wherein the machine learning model is generated based on a set of files and their metadata, wherein the set of files were classified based on their content.
4. The method of claim 1 , wherein the confidence level is determined based on a quality of the metadata.
5. The method of claim 1 , wherein the classification is also stored in directory information related to the file.
6. The method of claim 1 , wherein the classification is also stored in an organizational database accessible over an organizational network.
7. The method of claim 1 , wherein the metadata of the file includes classifications of neighbor files.
8. The method of claim 1 , wherein the metadata is enriched with information related to file storage metadata, including location ownership or a security level associated with the file.
9. A system for classifying a file locally at a computer, comprising:
the computer comprising a processor and memory;
a program for executing on the computer, wherein the program is configured to perform the following when executed by the computer:
extracting metadata from the file;
assigning a classification for the file by applying a machine learning model that was trained to classify files based on the metadata; wherein the machine learning model is provided by a remote server to classify files locally at the computer;
determining a confidence level representing an accuracy of the classification;
wherein if the confidence level is below a threshold value analyzing the content of the file to assign a classification for the file based on the content;
storing the assigned file classification; wherein the classification is stored as metadata in the file; and
wherein the assigned classification is used to permit or prevent access to the file by a user of the computer;
wherein at least one classification requires a second factor authentication by the user, wherein the user is required to enter a onetime code received in an SMS or Email in addition to entering a password.
10. The system of claim 9 , wherein the metadata is enriched with information related to file creator metadata or file editor metadata, including risk levels assigned to them.
11. The system of claim 9 , wherein the machine learning model is generated based on a set of files and their metadata, wherein the set of files were classified based on their content.
12. The system of claim 9 , wherein the confidence level is determined based on a quality of the metadata.
13. The system of claim 9 , wherein the classification is also stored in directory information related to the file.
14. The system of claim 9 , wherein the classification is also stored in an organizational database accessible over an organizational network.
15. The system of claim 9 , wherein the metadata of the file includes classifications of neighbor files.
16. The system of claim 9 , wherein the metadata is enriched with information related to file storage metadata, including location ownership or a security level associated with the file.
17. A non-transitory computer readable medium comprising an executable program configured to perform a method of claim 1 .