Method and System for Enhanced Content Recommendation

摘要

Method, system, and programs for providing content recommendation are disclosed. A first set of candidate content items may be generated based on a user profile, and a second set of candidate items may be generated based on the likelihood that the user will click a corresponding candidate content item in the second set. The candidate content items in the first and second sets may be ranked together using a learning model and presented to the user as content recommendations based on their rankings. The likelihood that the user will click a given candidate content item in the second set may be estimated based on similarities between the given content item and content items related to the given content item. Such a similarity may be computed based on activities performed by users who have viewed both the given content item and a related content item.

主权项

1. A method, implemented on a machine having at least one processor, storage, and a communication platform connected to a network, for recommending, to a user, content items, the method comprising:
obtaining a user profile characterizing interests of the user; generating a first set of candidate content items based on the user profile; generating a second set of candidate content items based on a likelihood that the user clicks a corresponding candidate content item in the second set, wherein each likelihood is estimated based on similarities between the candidate content items in the second set and one or more content items that were previously viewed by the user; ranking each of the candidate content items in the first set and the second set; and providing, based on the rankings, the candidate content items in the first and second sets as content recommendations to the user.