Hybrid Automatic Multistage Recommendation Techniques for Social Media Networks

The researchers on web recommendations on social media networks only consider similarity of web user's behavior and similarity of resources used by web user but not consider web user tagging similarity index. Collaborative filtering method facing sparsity problem, content filtering method not able to predict future web user interested web documents while recommending web pages, and web usage mining also not able to recommend newly added web documents to web users. In this research paper we propose hybrid multistage recommendation techniques for social media networks to resolve problems faced by previous recommendation techniques. Similarities matrix are measured using various performance evaluation metrics. Our proposed method outperforming when compared with traditional content filtering or collaborative filtering methods.