CPC G06F 16/38 (2019.01) [G06F 16/313 (2019.01); G06F 16/337 (2019.01); G06F 16/9024 (2019.01); G06F 40/205 (2020.01); H04L 51/42 (2022.05)] | 20 Claims |
1. A method comprising:
identifying, via a computing device, an inbox of a user comprising a set of messages;
parsing, via the computing device, each identified message in the set;
identifying, via the computing device, based on said parsing, message data and metadata for each message;
identifying, via the computing device, a criteria associated with a content recommendation, said criteria corresponding to a respective content type;
analyzing, via the computing device, the message data and metadata and identifying a number of content types including said respective content type based on said analysis;
generating, via the computing device, based on said analysis, a confidence graph comprising an entry for each content type of the number of identified content types, said confidence graph comprising a score for each identified content type, the score for said respective content type being based on the number of messages in the set mapping to said criteria corresponding to said respective content type, said respective content type's score indicating a degree of confidence in the user's interest in said respective content type; and
generating, via the computing device, an interest profile for the user based on said generated confidence graph comprising the score for said respective content type, the interest profile comprising information indicating an interest of the user in said respective content type.
|