A Recommender System for Digital Newspaper Readers Based on Random Forest

Abstract

In this research, the potential of machine learning methods based on decision trees (DT) and Random Forest (RF) models developed in the context of classifying readers of a digital newspaper. For this purpose, the number of visits of users to each section of the newspaper in a 3-month interval has been taken into account. The models of DT and RF developed in this paper classify the profiles of readers who access the journal with an accuracy of 98.07% and AUC value of 99.27%, thus demonstrating that it serves as a valid tool for making strategic and operational decisions when creating, manage and present content in the user – website interaction. © 2022, The Author(s), under exclusive license to Springer Nature Switzerland AG.