An Unsupervised Topic-driven New York Times Recommendation System

• Published in: International Symposium on Innovations in Intelligent Systems and Applications (Online) pp. 1 - 6
• Main Author: Petrusel, Mara-Renata
• Format: Conference Proceeding
• Language: English
• Published: IEEE 08.08.2022
• Subjects: Clustering; Clustering algorithms; Intelligent systems; kNN; Machine learning algorithms; New York Times Articles; Portals; Recommendation Systems; Recommender systems; Technological innovation
• ISSN: 2768-7295
• DOI: 10.1109/INISTA55318.2022.9894164

Recommendation systems are viral tools that make users' lives easier in terms of decisions and actions. In the online news portals, users are interested in specific domains and want to take a deep dive into the hottest topics of the moment. In this context, a content-based recommendation system helps identify the most relevant articles, satisfying their current needs and curiosities.The proposed New York Times recommendation system aims to suggest news articles for the New York Times readers, using an unsupervised machine learning-based approach. The k-Means clustering algorithm detects clusters of articles grouped based on similar topics or subjects. Then, the k Nearest Neighbors (kNN) algorithm is applied to determine the most similar articles to the target one based on the previously computed clusters of topics.The good results obtained for both the clustering and the recommendation processes in the conducted numerical experiments show that the proposed system is an excellent solution to the information overload problem in the online news portals.