Emerging Research Challenges and Opportunities in Computational Social Network Analysis and Mining by Nitin Agarwal & Nima Dokoohaki & Serpil Tokdemir

Author:Nitin Agarwal & Nima Dokoohaki & Serpil Tokdemir
Language: eng
Format: epub
ISBN: 9783319941059
Publisher: Springer International Publishing

3 Reputation Model

The focus of the proposed reputation system is on how to derive a reputation score for each node or user in OSNs; an individual user or an organization, who has an account. In this work, we consider a graph where an entity is a vertex and the social relationship between two entities; similar to other relevant studies (e.g., [23, 24, 30]). That is, if two entities are connected with an edge, this means that the two entities are friends with each other. Nonetheless, the degree of trust one entity has towards the other entity can vary, implying the trust relationship is subjective and asymmetric. Our reputation model in OSNs aims at assigning a reputation score for each node. This value is weighted so that users with different volumes of interactions can be compared with each other. Two major inputs are considered in this calculation: The interactions between the user as a node with other users or nodes. In graph models those interactions are modeled as edges. Unlike classical approaches, our model differentiates the thickness of the edge or the relation between friends through the volume of interactions between those friends. This weighted edge is bi-directional where the value in one direction is typically different from the other direction. One of the motivations in this scope is to be able to distinguish at a finder detail the weight of the relation between friends more than the classical binary classification (i.e., as a friend or not). We will explain our weighted-based model and how is this going to be utilized in different applications.

Download

Copyright Disclaimer:
This site does not store any files on its server. We only index and link to content provided by other sites. Please contact the content providers to delete copyright contents if any and email us, we'll remove relevant links or contents immediately.