An Intelligent and Personalized News Recommendation Model using Artificail Bee Colony Optimization-based Reinforcement Learning

Abstract

A personalized news recommendation system is a crucial and engaging area aimed at tailoring news content according to individual reading habits. Numerous research studies have been conducted over the years, resulting in significant advancements in predicting user behavior and enhancing user experience. Despite this progress, several challenges remain unresolved and demand further investigation. Traditional news recommendation systems follow a conventional approach to content delivery, which often lacks the adaptability required for long-term sustainability. This paper introduces a personalized, enriched news-feeding system that leverages a reinforcement learning approach integrated with the Artificial Bee Colony Optimization (ABCO) algorithm. The system recommends news based on a personalized wish list and commonly researched attributes in news content analysis. An agent node generates a list of recommended news items by gathering personal user information, while the environment evaluates and rewards recommendations based on the user’s reading patterns. The agent node is developed using the ABCO algorithm, which facilitates the generation of enriched and relevant news suggestions. The performance of the proposed system, evaluated in terms of recommendation accuracy, is based on personalized data and has been compared against existing models. Experimental results show that the proposed system achieves an accuracy rate of 97.5% in news recommendation.