Recommendation method and device based on probability sampling, equipment and medium

Abstract

The invention discloses a recommendation method and device based on probability sampling, equipment and a medium. The method comprises the steps of obtaining a behavior log; generating an initial sample set according to the behavior log; sampling each object in an initial sample in the initial sample set according to a probability sampling algorithm to obtain a training sample corresponding to theinitial sample; obtaining a training sample set according to the training samples corresponding to the plurality of initial samples; training a recommendation model according to the training sample set, and obtaining a word embedding vector corresponding to each object in the training sample set based on the recommendation model; obtaining a candidate object set, and obtaining a word embedding vector corresponding to each candidate object in the candidate object set based on the recommendation model; and extracting a recommended object set from the candidate object set according to the word embedding vector of each object in the training sample set and the word embedding vector corresponding to each candidate object in the candidate object set. Objects with more diversity can be recommended, and the user viscosity is improved.

Description

technical field [0001] The present invention relates to the field of recommendation, in particular to a recommendation method, device, equipment and medium based on probability sampling. Background technique [0002] The media service platform in the prior art often involves recommending potential objects of interest to users, for example, recommending news objects, audio and video objects, or graphic objects that users are interested in. The recommendation basis is the user's historical records, that is, recommending objects of interest similar to the historical records to the user according to the user records. [0003] In the prior art, training samples are usually constructed offline, a recommendation model is trained according to the training samples, and objects of interest are recommended for users by relying on the recommendation model for online recommendation. The offline construction of training samples means extracting the objects selected by the user as trainin...

AI Technical Summary

Problems solved by technology

[0004] In order to solve the technical problem of low recommendation accuracy caused by the low quality of training samples in the prior art, the embodiment of the present invention provides a recommendation method, device, equipment and medium based on probability sampling

Images