Intelligent file recommendation engine
Methods and systems for recommending files to users are described herein. Files may be recommended to a user within a file sharing service. A recommender system may intelligently recommend files to users according to their preferences through machine learning. In addition, a recommender system may recommend files based on what is popular within a group to which the user belongs. The recommendations may be adjusted based on user interaction with one or more recommended files.
1. A computer-implemented method, comprising:
determining, by a computing device comprising a processor, user behavior data corresponding to a user within a file sharing service, wherein the user is part of a group of users;
generating, by the computing device, one or more feature vectors based on the user behavior data, and based on group behavior data corresponding to actions taken by the group of users within the file sharing service;
generating, by the computing device, using a first recommender model and based on the one or more feature vectors, a first set of recommended files for the user;
causing, by the computing device, to display the first set of recommended files to the user;
determining that user interaction with the first set of recommended files fails to satisfy a threshold; and
causing to display a second set of recommended files for the user, wherein the second set is generated using a second recommender model.
2. The computer-implemented method of claim 1 , further comprising:
after determining that the user interaction with the first set of recommended files fails to satisfy the threshold,
generating a modified first recommender model by modifying training parameters of the first recommender model; and
generating the second set of recommended files for the user using the modified first recommender model.
3. The computer-implemented method of claim 1 , wherein determining that the user interaction with the first set of recommended files fails to satisfy the threshold comprises determining an amount of time the user spends viewing one or more files of the first set of recommended files.
4. The computer-implemented method of claim 1 , wherein determining that the user interaction with the first set of recommended files fails to satisfy the threshold comprises determining whether one or more files of the first set of recommended files was edited by the user.
5. The computer-implemented method of claim 1 , wherein the user behavior data comprises information corresponding to files viewed by the user.
6. The computer-implemented method of claim 1 , wherein the user behavior data comprises information corresponding to files modified by the user.
7. The computer-implemented method of claim 1 , wherein the group behavior data comprises information indicating files shared within the file sharing service by one or more users of the group of users.
8. The computer-implemented method of claim 1 , wherein the group behavior data comprises information indicating files viewed within the file sharing service by one or more users of the group of users.
9. A system comprising:
a server and a user device,
wherein the server comprises:
one or more processors and memory, configured to:
determine user behavior data corresponding to a user within a file sharing service, wherein the user is part of a group of users;
generate one or more feature vectors based on the user behavior data, and based on group behavior data corresponding to actions taken by the group of users within the file sharing service;
generate, by a first recommender model and based on the one or more feature vectors, a first set of recommended files for the user;
cause to display the first set of recommended files to the user device;
determine that user interaction with the first set of recommended files fails to satisfy a threshold; and
cause to display a second set of recommended files for the user, wherein the second set is generated using a second recommender model.
10. The system of claim 9 , wherein the one or more processors and memory are further configured to:
after determining that the user interaction with the first set of recommended files fails to satisfy the threshold,
generate a modified first recommender model by modifying training parameters of the first recommender model; and
generate the second set of recommended files for the user using the modified first recommender model.
11. The system of claim 9 , wherein determining that the user interaction with the first set of recommended files fails to satisfy the threshold comprises determining an amount of time the user spends viewing one or more files of the first set of recommended files.
12. The system of claim 9 , wherein determining that the user interaction with the first set of recommended files fails to satisfy the threshold comprises determining whether one or more files of the first set of recommended files was edited by the user.
13. The system of claim 9 , wherein the user behavior data comprises information corresponding to files viewed by the user.
14. The system of claim 9 , wherein the user behavior data comprises information corresponding to files modified by the user.
15. A non-transitory machine-readable medium storing instructions, that when executed by one or more processors, cause the one or more processors to:
determine user behavior data corresponding to a user within a file sharing service, wherein the user is part of a group of users;
generate one or more feature vectors based on the user behavior data, and based on group behavior data corresponding to actions taken by the group of users within the file sharing service;
generate, by a first recommender model and based on the one or more feature vectors, a first set of recommended files for the user;
cause to display the first set of recommended files to a user device associated with the user;
determine that user interaction with the first set of recommended files fails to satisfy a threshold; and
cause to display a second set of recommended files for the user, wherein the second set is generated using a second recommender model.
16. The non-transitory machine-readable medium of claim 15 , wherein the group behavior data comprises information indicating files shared within the file sharing service by one or more users of the group of users.
17. The non-transitory machine-readable medium of claim 15 , wherein the user behavior data comprises information corresponding to files modified by the user.
18. The non-transitory machine-readable medium of claim 15 , wherein the group behavior data comprises information indicating files viewed within the file sharing service by one or more users of the group of users.