Systems And Methods For Recommendation System Based On Implicit Feedback
Systems and methods are described for providing a recommendation system for a vehicular content distribution network. A static recommendation list can be generated based on travel characteristics stored in a server. An efficiency threshold can be calculated or provided that sets the point at which the static recommendation should no longer be used for recommendations and the recommendation system can be used. Data is gathered of each user's preference based on implicit feedback, and the data is then analyzed to calculate an efficiency level of the recommendation system operating with gathered data for a user. The recommendation system is automatically used for a specific user when the efficiency level meets or exceeds the efficiency threshold.
1 . A method for providing a recommendation system for a vehicular content distribution network, comprising:
providing a server having a processor and memory, wherein the server is communicatively coupled to an in-vehicle network for distributing content to a plurality of users;
computing a static recommendation list using the processor based on travel characteristics stored in the memory;
computing an efficiency threshold at which the static recommendation should no longer be used for recommendations using the processor;
gathering data on each user's preferences based on implicit feedback and storing the data in the memory; and
analyzing the gathered data using the processor to calculate an efficiency level of the recommendation system operating with gathered data using a processor for a user, and automatically switching from the static recommendation list to a recommendation system for that user when the efficiency level meets or exceeds the efficiency threshold.
2 . The method of claim 1 , wherein the implicit feedback comprises interaction of the user with an in-flight entertainment system.
3 . The method of claim 2 , wherein the interaction comprises one or more of (i) the user selecting a piece of content, (ii) the user reviewing details of a piece of content, (iii) the user liking a piece of content, (iv) the user adding a piece of content to a playlist, and (v) the user skipping a piece of content.
4 . The method of claim 1 , wherein the efficiency threshold is based at least in part on flight characteristics stored in the memory.
5 . The method of claim 4 , wherein the flight characteristics comprises at least one of (i) a length of the flight, (ii) an amount of content available on the flight, (iii) a type and/or diversity of content available on the flight, and (iv) feedback from passengers on prior flights.
6 . The method of claim 4 , wherein the efficiency threshold varies for different flights.
7 . The method of claim 1 , wherein the travel characteristics comprise user demographics.