Research on Distributed Machine Learning Model for Predicting Users’ Interest by Acquired Web Contents Similarity

This paper discusses and proposes a method for predicting and analyzing the current user interests based on their characteristics including their own recently acquired web content, and it is used. The traditional methods of using cookies on browsers and capturing users’ web activities cannot be used in the future for privacy reasons. It is difficult for anyone other than the users themselves to obtain accurate web activities. Thus, all methods based on the identification of individual users, such as user similarity comparisons between users, are no longer available. This paper focuses on the web content information obtained by users of each web service. Assuming that the web content acquired by each user contains his or her current interests at that point in time, a combinable machine learning model named the fog model is acquired from web content that is similar to these features. The fog models are then combined in the same manner along the most recent web activities. The combined fog models include the features of each web service, which can be considered to be equivalent to the features obtained by accumulating and analyzing the entire content. As an evaluation of the proposed method, the performance of the proposed method is compared to that of the conventional method using cookies on the data from targeted advertisements. © 2023, The Author(s), under exclusive licence to Springer Nature Singapore Pte Ltd.