Abstract
Recently, social networks have shown huge potential in terms of collaborative web services and the study of peer influence as a result of the massive amount of data, datasets, and interrelations generated. These interrelations cannot guarantee the success of online social networks without ensuring the existence of trust between nodes. Detecting influential nodes improves collaborative filtering (CF) recommendations in which nodes with the highest influential capability are most likely to be the source of recommendations. Although CF-based recommendation systems are the most widely used approach for implementing recommender systems, this approach ignores the mutual trust between users. In this paper, a trust-based algorithm (TBA) is introduced to detect influential spreaders in social networks efficiently. In particular, the proposed TBA estimates the influence that each node has on the other connected nodes as well as on the whole network. Next, a Friend-of-Friend recommendation (FoF-SocialI) algorithm is addressed to detect the influence of social ties in the recommendation process. Finally, experimental results, performed on three large scale location-based social networks, namely, Brightkite, Gowalla, and Weeplaces, to test the efficiency of the proposed algorithm, are presented. The conducted experiments show a remarkable enhancement in predicting and recommending locations in various social networks.
Highlights
A noticeable focus in research activities is devoted to answering the question of how to efficiently detect the most influential nodes in social networks
Motivated by the significant influence of these research directions in many application areas, we followed efforts introduced in previous works and presented the following contributions: 1. We investigated the role of influential nodes in the recommendation process and in location recommendation
We propose an efficient algorithm (TBA) for discovering influential nodes in online social networks and enhancing the location recommendation services
Summary
A noticeable focus in research activities is devoted to answering the question of how to efficiently detect the most influential nodes in social networks. Hangal et al proposed an algorithm for enhancing recommendations by considering both weighted and directed edges in the social graph [16]. They proved experimentally that their proposed algorithm showed an improvement in recommendation accuracy by approximately 35%. We extended the algorithm [17] by proposing a novel Friend-of-Friend (FoF) location recommendation algorithm that computes location recommendations employing FoF relationship information The focus of this investigation was to study and compute trust in social networks rather than a study of the network literature.
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
