A personalized content recommendation method for teenagers

Abstract

The association of the present invention has developed a personalized content recommendation method for young people, including the following steps, S1, collecting historical browsing behaviors of users for each recommended content, and using it as a training set for a neural network model; S2, constructing a neural network Network model; S3, preprocessing the training set, and incorporating the preprocessed training set into the neural network model, obtaining the inclusion result, calculating the mean square error of the inclusion result, and taking the result of minimizing the mean square error as The optimization goal is to perform model training on the neural network model; S4, select a user, and when recommending content to the user, perform recommendation scoring on the content recommended to the user. The advantages are: improve the matching degree between recommended content and user's age; increase the exposure of content that matches the user's age characteristics in addition to the user's historical interest, and broaden the vision of young users; , to avoid overfitting the user's historical preferences and forming an information cocoon.

Description

technical field [0001] The invention relates to the field of personalized recommendation algorithms, in particular to a personalized content recommendation method for teenagers. Background technique [0002] With the rapid development of information technology and network technology, global information has experienced explosive growth, and massive data is presented in front of people. While enjoying rich information resources, people are also troubled by how to obtain the part of information that is really useful to them. . In the face of this "data overload" problem, two tools, search engines and recommendation engines, have emerged to help understand users' information needs; among them, search engines are oriented to users' explicit intentions, that is, users have clear information acquisition needs; and recommendation The engine is oriented to the user's implicit intention, that is, the user does not have a clear need for information acquisition. Among them, recommenda...

AI Technical Summary

Problems solved by technology

In addition, due to the development of machine learning technology in recent years, a number of personalized recommendation technologies based on neural network models have emerged, such as factorization machines, Wide&Deep neural networks, etc.; however, in algorithm design, no specific considerations have been made for the youth user group Blindly pursuing the rise of CTR will lead to the gradual narrowing of vision, and the long-term formation of information cocoons will stifle the creativity of young people

Images