Indian Journal of Science and Technology
Year: 2024, Volume: 17, Issue: 11, Pages: 1003-1015
Original Article
Anil Kumar1,2*, Sonal Chawla3
1Research Scholar, Department of Computer Science and Applications, Panjab University, Chandigarh, India
2Assistant Professor, Department of Computer Science, G.G.D.S.D College, Hariana, Hoshiarpur, Punjab, India
3Professor, Department of Computer Science and Applications, Panjab University, Chandigarh, India
*Corresponding Author
Email: [email protected]
Received Date:16 January 2024, Accepted Date:09 February 2024, Published Date:29 February 2024
Objectives: This study presents an efficient approach "Hybrid Sentiment-based Collaborative Architecture" to enhance book recommendation systems. This novel approach integrates sentiment analysis methodologies that encompass Lexicon-based and Deep Learning-based techniques, in conjunction with Collaborative Filtering to offer a more personalized recommendation experience. Methods: This study outlines the methodology for comparing and analyzing various Collaborative Filtering and sentiment analysis techniques to identify an optimal combination. A public dataset “Amazon book review dataset” is employed for the experimental work. In this experimental study, 75% of the dataset serves as the training dataset, and 25% is designated as the testing set. Evaluation of the proposed hybrid approach involves standard metrics such as accuracy, precision, recall, and F1-Score. Findings: The proposed hybrid architecture overcomes the drawbacks of traditional recommendation systems by using users' past behavior and preferences through Collaborative Filtering, and incorporating sentiment analysis to understand the emotional tone of reviews. Results and conclusions derived from evaluating the effectiveness of the hybrid architecture in book recommendations provide insights into potential advancements in recommendation system paradigms. The proposed approach improves the recognition accuracy by 80.95% as compared to other existing systems in literature and possible hybridizations. The proposed methodology demonstrates significant enhancements in precision and F1-Score. Novelty: The proposed framework employs numerical ratings and sentiments to prognosticate recommendations, with the ultimate suggestion incorporating the relative significance of product sentiments and numerical ratings using the Collaborative Filtering technique and sentiment analysis technique incorporating Lexicon-based and Deep Learning-based techniques.
Keywords: Recommendation Systems, Book Recommendation System, Machine Learning, Sentiment Analysis, Deep Learning
© 2024 Kumar & Chawla. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Published By Indian Society for Education and Environment (iSee)
Subscribe now for latest articles and news.