Recommender system and method for generating implicit ratings based on user interactions with handheld devices

A method for generating recommendations, comprising

• providing an item of a particular type to a handheld device having an application for engaging in a repetitive activity with items of the particular type, wherein the repetitive activity comprises displaying or playing items of the particular type to a user;
• generating a history of user interaction with the provided item, wherein a user interaction comprises an instance of a user causing the application to display or play the provided item to the user and duration of the display or play, wherein each user interaction occurs during standalone operation of the handheld device disconnected from a network;
• wherein the history of user interactions is transformed into data pertaining to normalized time spent interacting with the provided item;
• uploading the history of user interactions to a network recommender;
• transforming the history into an implicit rating of the provided item, wherein the history of user interactions with the provided item may be used to create more accurate statistical profiles the rating comprising predicted ratings for a user for a plurality of items not rated by the user, having a measure of confidence in the prediction and a rationale for the prediction;
• using the implicit rating of the provided item to generate recommendations of other items of the particular type;
• and
• wherein an implicit rating is generated in accordance with the relationship:
• rating(item)=total interaction time(item)/size(item).

A method for generating recommendations, comprising:

• providing an item of a particular type to a handheld device having an application for engaging in a repetitive activity with items of the particular type, wherein the repetitive activity comprises displaying or playing items of the particular type to a user;
• generating a history of user interaction with the provided item, wherein a user interaction comprises an instance of a user causing the application to display or play the provided item to the user and duration of the display or play, wherein each user interaction occurs during standalone operation of the handheld device disconnected from a network;
• wherein the history of user interactions is transformed into data pertaining to normalized time spent interacting with the provided item;
• uploading the history of user interactions to a network recommender;
• transforming the history into an implicit rating of the provided item, wherein the history of user interactions with the provided item may be used to create more accurate statistical profiles the rating comprising predicted ratings for a user for a plurality of items not rated by the user, having a measure of confidence in the prediction and a rationale for the prediction;
• using the implicit rating of the provided item to generate recommendations of other items of the particular type;
• and
• wherein an implicit rating is generated in accordance with the relationship:
• rating(item)=[total interaction time(item)/size(item)*exp(−damping coefficient]*(date−time acquired).