Predicting Citation Dynamics and Mapping Research Trends in Nanocellulose: A Bibliometric and Machine Learning Approach (2021–2025)

Abstract

This study aims to map research trends and predict citation dynamics in nanocellulose research within the materials science domain during 2021–2025. A quantitative bibliometric approach was employed using metadata retrieved from the Scopus database, followed by network visualization with VOSviewer and advanced data analysis using Python and machine learning techniques. A total of 2,971 publications were analyzed to identify publication patterns, collaboration networks, thematic evolution, and citation behavior. The results show that China dominates publication output and funding, while key journals and authors form highly interconnected citation networks. Topic modeling reveals emerging research fronts in biomedical hydrogels, nanocomposite films, and sustainable processing. Citation prediction using regression-based machine learning achieved moderate performance (R² = 0.23), indicating potential for early impact estimation. This study concludes that integrating bibliometrics with machine learning provides a comprehensive and predictive perspective on the evolving landscape of nanocellulose research and can support strategic research planning and policy decisions