|
Volume 10, Number 1, January 2025 , pages 91-109 DOI: https://doi.org/10.12989/acd.2025.10.1.091 |
||
|
An intelligent hybrid recommendation system for enhancing viewer experience |
||
Kulvinder Singh, Sanjeev Dhawan and Manoj Yadav
|
||
| Abstract | ||
| We introduce HybridRecSys, a hybrid recommendation system that integrates collaborative filtering (CF) using enhanced Singular Value Decomposition (SVD) and advanced content-based (CB) filtering techniques enriched by Natural Language Processing (NLP). The proposed system addresses critical challenges such as sparsity and cold start by leveraging a dual approach: explicit ratings for strong user profiling and implicit preferences derived from content and genre analysis. Novel contributions include the application of weighted cosine similarity alongside RBF and cosine similarity, significantly improving similarity metrics. Experimental validation on IMDb and Netflix datasets demonstrates superior performance, with HybridRecSys achieving RMSE and MAE scores of 0.6991 and 0.6987 on IMDb, and 0.2364 and 0.2357 on Netflix, respectively. The system outperforms existing methods by efficiently addressing sparsity and cold start challenges, ensuring highly personalized and accurate recommendations. | ||
| Key Words | ||
| collaborative filtering; content-based filtering; enhanced SVD; HybridRecSys; temporal dynamics | ||
| Address | ||
| Kulvinder Singh, Sanjeev Dhawan and Manoj Yadav: Department of Computer Science & Engineering, University Institute of Engineering & Technology (U.I.E.T), Kurukshetra University, Kurukshetra, Haryana, India | ||
