Multi-task television program recommendation method integrating user click behaviors as well as user interests and preferences

Abstract

The invention discloses a multi-task television program recommendation method integrating user click behaviors as well as user interests and preferences, which improves the television program recommendation effect through integrating user clicking and multi-task learning of user interests and preferences in television program recommendation. The multi-task television program recommendation methodcomprises the steps of: establishing a clicking model of television programs by a user from historical clicking program data of the user based on an attention network fusion neural network factorization algorithm; establishing an interest and preference model of the user for the programs from the user interest and preference data based on the attention network fusion neural network factorization algorithm; and finally, comprehensively considering the click behaviors of the user and the preference behaviors of the user during recommendation, so that the recommended television programs can better meet the requirements of the user, and the satisfaction degree of the user is improved.

Description

technical field [0001] The invention relates to the field of multi-task TV program recommenders, in particular to a multi-task TV program recommendation method that integrates user click behavior and user interest preference. Background technique [0002] With the development of the Internet and information technology, in order to solve the problem of information overload, the research and development of recommender systems has become a hot spot, and many achievements have been made in both industry and academia. As the demand of TV users is getting higher and higher, the single-task recommendation system often only considers the user's click behavior, and does not consider whether the user likes the TV program after clicking, and the user's interest preference is particularly important in the recommendation process. , so the original recommendation system that only considered the user's click behavior is not able to meet the user's personalized recommendation needs. Furthe...

AI Technical Summary

Problems solved by technology

As the demand of TV users is getting higher and higher, the single-task recommendation system often only considers the user's click behavior, and does not consider whether the user likes the TV program after clicking, and the user's interest preference is particularly important in the recommendation process. , so the original recommendation system that only considered the user's click behavior is not able to meet the user's personalized recommendation needs

Images