Article recommendation method, device and equipment and computer readable storage medium

Abstract

The invention discloses an article recommendation method, which comprises the steps of analyzing a received article recommendation request to obtain request end identity information and a to-be-recommended target article; inputting the request end identity information and the target object into a target restricted Boltzmann machine collaborative filtering recommendation model obtained by trainingthrough a comparison divergence algorithm for increasing impulse items; utilizing the target restricted Boltzmann machine collaborative filtering recommendation model to obtain a historical score setcorresponding to the target request end according to the request end identity information, and performing scoring operation on the target article according to the historical score set to obtain a scoring result of the target request end about the target article; and generating a recommendation result of the target article about the target request end according to the scoring result, and outputtingthe recommendation result. By applying the technical scheme provided by the embodiment of the invention, the recommendation efficiency is greatly improved, and the article recommendation accuracy isimproved. The invention furthermore discloses an article recommendation apparatus and device, and a storage medium, which have corresponding technical effects.

Description

technical field [0001] The present invention relates to the field of computer application technology, in particular to an item recommendation method, device, equipment and computer-readable storage medium. Background technique [0002] Facing the increasingly prominent contradiction between information overload and personalized needs, the recommendation system has become an effective means to solve this problem. Among personalized recommendation systems, collaborative filtering algorithm is one of the most widely used recommendation techniques. [0003] The existing item recommendation method mainly adopts the collaborative filtering algorithm based on the traditional restricted Boltzmann machine (RBM) to realize the user's recommendation according to the similarity of interests between users and the user's interest preference. There are some defects in the collaborative filtering algorithm based on the traditional restricted Boltzmann machine: First, the collaborative filt...

AI Technical Summary

Problems solved by technology

There are some defects in the collaborative filtering algorithm based on the traditional restricted Boltzmann machine: First, the collaborative filtering algorithm based on the traditional restricted Boltzmann machine is easy to fall into local optimum in the application, which makes its application scope limited, and the The recommendation result is inaccurate; secondly, when the amount of data is too large, the number of iterations of the restricted Boltzmann machine will increase exponentially. After a very long learning and training process, the expected root mean square error requirement can be achieved, and the obtained Satisfactory data results

Therefore, when the traditional restricted Boltzmann machine is applied to the real-time recommendation system, in the face of a large amount of network data, it takes a long time to calculate the results, the recommendation efficiency is low, and the user's real-time experience is poor.

Images