A novel re-share recommender system for optimizing targeted content diffusion in social networks

Content diffusion is the process of spreading information through members of a social network, with numerous applications in various fields such as marketing and advertising. A significant challenge in this domain is the lack of alignment between content and user interests in existing solutions, which primarily focus on network topology. Additionally, previous methods, like those in [3], often overlooked the consideration of negative feedback in the diffusion process. Most past studies have concentrated on selecting seed nodes or maximizing diffusion reach, which results in less attention being paid to targeted content dissemination and preventing diffusion among non-target users. This paper presents a content diffusion recommendation system that considers target and non-target nodes. The proposed approach combines users' structural and behavioral features in the social network to optimize the selection of seed nodes and facilitate recommendations for content diffusion. The system directs content toward target users by optimizing pathways during diffusion, improving reach, and reducing costs. Initial results on an Instagram-derived dataset (1,204 nodes, 4,238 edges) indicate improved performance, achieving approximately 90% efficiency in reaching target nodes and reducing content diffusion among non-target nodes by 70-75% compared to general diffusion. Efficiency is defined as the ratio of activated target nodes to total targets. By combining structural and behavioral features, this system extends prior work like [3], enhancing precision and cost-effectiveness in targeted diffusion. • Proposed a re-share recommender system for targeted content diffusion. • Integrates structural & behavioral features for targeted diffusion. • Optimizes seed node selection in social networks • Achieves 90% target activation efficiency in Instagram network. • Reduces non-target diffusion by about 70–75%.