Optimizing the choice of influential nodes for diffusion on a social network

Abstract

Summary An optimization approach is taken to locating the optimal set of initial contacts in a social network to maximize the number of total network members reached by a message. It is assumed that initial contacts are costly and that the number of initial contacts must be minimized simultaneously with maximizing the total number of network members contacted. A bi‐objective probabilistic integer programming model is developed that assumes that actors are heterogeneous in the probability that they will pass messages along their ego networks. Considering the complexity of solving the proposed model, it reformulated as a pure integer programming model. The algorithm is illustrated by the analysis of message passing in a short‐message system (texting) among university students. Copyright © 2015 John Wiley & Sons, Ltd.

A hybrid algorithm based on community detection and multi attribute decision making for influence maximization

“…Liu et al proposed a bound linear approach to influence computation and influence maximization [38]. Kermani et al [1] proposed a multi objective mathematical program with linear objectives and constraints in search of the seed nodes in social networks. Since the proposed model was solved by an exact algorithm (CPLEX), they claimed that their model finds the optimal solution for influence maximization problem.…”

“…Information diffusion is a vast research domain and has expressed research interests of many fields, such as Physics, Biology, etc. [23], [35] [1]. For example, considering Facebook, where a user Sally updates her status or writes on a friend's wall about a new show in town that she enjoyed.…”

“…In this work, the focus is on the non-competitive case in which there is a decision maker attempting to spread a piece of information on a social network as much as possible. In this paper, a new algorithm to solve the influence maximization problem is proposed with three main features: (1) The proposed algorithm takes the topological features of the nodes into account. (2) The nodes' features are taking into consideration simultaneously by making use of one of the multi attribute decision making methods.…”

Social recommendation between urban vehicular social communities

“…In the social recommendation research area, much work has been done on one‐to‐one individual social recommendation while research on social recommendation between communities is still insufficient till now . In this paper, based on the existing efficient trust evaluation mechanism, the optimal solution of social recommendation is studied using an individual external trust and internal trust evaluation model, based on which effective and efficient information dissemination between urban vehicular social communities is performed. Specifically, in this paper, there are three main contributions which are listed as follows.…”