Testing Assessment of Group Collaborations in OSNs
Additional Document Info
Interactions of users in Online Social Networks (OSNs) offer interesting insights to a large spectrum of researchers from different domains. With the continuously increasing volume of data exchanged between those users, it is reasonable to think of methods to improve information accuracy and also protect users privacy. In this research we proposed a weighted-based approach to describe relations between users in OSNs. Users in OSNs interact with each other through comments, posts, pictures, videos, etc. Our model investigates relations between users as nodes in OSNs, their activities as products of those nodes, and interactions between the different nodes through those activities. We showed how our reputation model can be used in several applications, including friends recommendation system, referrals, credit reporting, spam detection, and networks evolution. In social networks, a clique is a group of users in which everyone is a friend to all other group members. Interactions between cliques members are studied in different networks for knowledge extraction. We introduced the concept of weighted cliques in comparison with classical cliques to provide better understanding of users interactions. We showed using a demo dataset how to extract knowledge using the proposed method and how is that different from classical clique methods. We also proposed a context-driven model for privacy assessment in OSNs. The goal is to semi-automate the process of deciding the visibility levels of users created posts or activities.