A Hybrid Recommendation System Based on the Supply Chain in Social Networks

Abstract

With the expansion of virtual social networks, finding and recommending appropriate and favorite information and items to users is one of the severe issues in their development. To this end, recommender systems predict and recommend interests based on past behavior reviews and user preferences. However, less research has been done on people to people in social networks, and it is still based on exploring communication and friendship circles, which is generally not desirable for specialized users. Social networks include a variety of entities such as individuals, businesses, companies, and technical communications that also contain a variety of information related to the supply chain interaction, such as industries, functions, and communications between them and users. This paper provides a recommendation system framework for recommending people to people in social networks based on supply chain interactions. For this purpose, it has presented five hybrid methods based on artificial neural networks and fuzzy strategies to provide better and more accurate recommendations than basic methods. Eventually, a case study was conducted on the LinkedIn social network to show the improvements in applying this new approach to primary methods. In this regard, seven specific evaluation criteria of recommender systems have been used.