Media content rankings for discovery of novel content

摘要

A content provider system ranks media content items with respect to a particular user based on selection scores determined for each of the media content items. The selection scores may be determined using a particular model that calculates a predicted selection score based on feature values associated with the content item with respect to the particular user. The feature values may indicate properties of the media content item, the particular user, or the particular user's relationship with the content item, including information about the novelty of the media content item with respect to the user. The particular model may be trained with sample user consumption data points that represent various combinations of media content items and users. The data point information evaluated during the training of the particular model may cause the model to assign higher selection scores to content items that are novel in particular ways.

主权项

1. A method comprising:
in a data processing system configured to generate personalized rankings of a first plurality of media content items in a catalog: determining a plurality of user consumption values, wherein each user consumption value of the plurality of user consumption values is associated with a user consumption data point of a plurality of user consumption data points, and wherein each user consumption data point is associated with a user and a media content item in a second plurality of media content items; determining, for each user consumption data point of the plurality of user consumption data points, a plurality of feature values associated with the user consumption data point, wherein the plurality of feature values include a particular feature value indicating whether the user associated with the user consumption data point consumed the media content item associated with the user consumption data point within a particular amount of time; training a selection value prediction model based on each feature value of the plurality of feature values associated with each user consumption data point of the plurality of user consumption data points; for each media content item of the first plurality of media content items, determining, based on the selection value prediction model, a predicted selection score for an input user and the media content item; ranking the first plurality of media content items in the catalog for the input user based on the predicted selection score for each media content item in the first plurality of media content items; wherein the method is performed using one or more processors.