Multi-Level Multiple Learning-Based Recommendation (Mmlr) Model For Youtube Recommendation And Security Enhancement

Abstract

As many entertaining social media factors are evolving around online and YouTube is the topmost entertainer. In YouTube, the anchors' list increases day-by-day making the selection tough for the viewers to select their favorite anchors and follow them. To avoid these confusing circumstances and to effectively choose the right anchor an efficient recommendation system is required. The main aim of the paper is to produce an effective, secured data transaction over streaming platforms. Generally, the user’s preference and creator's preference changes over time. In the previous system, there were algorithms to study only viewer’s choices and prefer contents according to them. In the proposed system a deep learning and preference matching on both the anchors' preference and user’s preference is made manually and automatically. A novel ‘Multi-level Multiple learning-based recommendation (MMLR)’ model for YouTube recommendation is proposed. Further, the system concentrates on the data transaction and its security. Several optimized techniques such as cuckoo hashing, elliptic curve cryptographic algorithm (ECC), fault tolerance, and anonymity maintenance to provide secure data transaction on cloud streaming platforms. To evaluate the performance of the proposed model, the paper initializes experiments on real datasets, and the result proves that the system outperforms all other recommendation models.