A Dynamic and Context-Aware Social Network Approach for Multiple Criteria Decision Making Through a Graph-Based Knowledge Learning

A Dynamic and Context-Aware Social Network Approach for Multiple Criteria Decision Making Through a Graph-Based Knowledge Learning

Alessandro Di Stefano (University of Catania, Italy), Marialisa Scatà (University of Catania, Italy), Aurelio La Corte (University of Catania, Italy) and Evelina Giacchi (University of Catania, Italy)
DOI: 10.4018/978-1-5225-2814-2.ch004


Complexity and dynamics characterize a social network and its processes. Social network analysis and graph theory could be used to describe and explore the connectedness among the different entities. Network dynamics further increases the complexity, as each entity with its personal knowledge, cognitive and reasoning capabilities, thinks, decides and acts in a social network, characterized by the heterogeneity of nodes and ties among them. Social network analysis becomes critical to the decision-making process, where a network node will consider both its personal knowledge and the influences received from its neighbors. Network dynamics and the node's context-awareness affect the relationships among criteria, modifying their ranking in a multiple criteria decision-making process, and hence the decision itself. Thus, the main aim has been to model the decision-making process within a social network, considering both context-awareness and network dynamics. Moreover, we have introduced a process of knowledge-transfer, where the criteria are represented by the knowledge-related values.
Chapter Preview

1. Introduction

One of the most important challenges of future ICT is to develop innovative methodologies, tools and algorithms to extract knowledge from several and heterogeneous sources, linked to multiple types of data. The Big Data revolution allows to look at many perspectives of single or correlated aspects, and it raises the issue of unifying these heterogenous resources. The challenge is to combine relevant information in the most efficient way and obtain a mining process, leading to a comprehensive understanding and a real knowledge extraction. The extraction of knowledge is strictly related to the value it can generate in a network, other than its rate of growth. A key aspect in this future ICT approach is based on adopting a bio-inspired approach applied to social networks, making nodes increasingly more smart and human, by introducing cognitive modules and making them able to decide and make inferences using context-aware strategies according to the specific context. Nodes are part of a complex social network, whose dynamics, influences and contagion processes affect the decision process of the single nodes and communites they belong to. The ubiquitous and dynamic nature of the network requires smart entities, which can decide using context-aware strategies according to the specific context (Scatà et. al., 2014).

To better analyze, in a deeper way, each process in a social network, there is a need to mine knowledge from variours sources, studying the phenomena occuring within a social network. Each process is characterized by a number of entities, which gives a contribute in order to establish the entire path that forms the whole system. Each entity, interacting through its relationships in the social network, is inevitably influenced in its opinion and actions as a consequence (Asavathiratham et. al. 2001; Grabisch & Rusinowska 2010; Barjis et. al., 2011; Pachidi et. al., 2014). As reported in (López-Pintado, 2008), “individual decisions are often influenced by the decisions of other individuals”. Without considering any interaction, nodes easily rank criteria in terms of importance following an individual cognitive model (Korhonen & Wallenius 1997). Therefore, each individual cannot be considered as an isolated entity, since the behavior of each entity is the result of the interactions between its preferences and the dynamic nature of relationships within the social network, impacting every individual’s decision, inserting it in a complex and dynamic social perspective (Pentland 2014). Social phenomena are characterized by the influence, which is related to every process inside the network, such as innovation diffusion, cultural events, and each flow which involves entities. As a consequence of the influence exercised by the nodes in the network, the preferences of each node could change during the decision-making process bringing to different decisions over time.

As highlighted by a vast literature, social networking plays a key role in the decision-making process (Kempe et. al., 2015; Anagnostopoulos et. al., 2015). Some works have focused on the importance of influence inside social networks. In (kempe et. al., 2015), the authors have faced the issue of influence maximization in viral marketing aplications. Recently, in (Anagnostopoulos et. al., 2015), it has been presented a model for the diffusion of competing alternatives in a social network, where nodes decide among different alternatives.

Complete Chapter List

Search this Book: