Βιβλιογραφική επισκόπηση συστημάτων κοινωνικής σύστασης

Abstract

Social networks have become a source of clear and also implicit information for their users. The amount of information with commercial content that can now be found on social networks is vast. In this area, some research has focused on whether automated recommendation systems can be implemented, taking into account the user interactions on social networks. These systems are called social recommendation systems (SRs). As users spend more and more time on social networks and share valuable information about themselves and their relationships, the use of this information helps to optimize the mechanisms for creating automated proposals in the above systems. The literature review in the context of this dissertation highlighted the need to address some problems which remain open challenges. First of all, raw data requires further attention, such as how to extract the most useful words in a user's comments, what concepts or emotions are associated with those words, and what suggestions should be made about those user's concepts or emotions. The problem of ambiguity, and the optimization of filters, either on the basis of content or on the basis of the relationships that users form on social networks, also remains unresolved. This work highlights different methods that either attempt to model how an individual's social relationships can influence their decisions to buy a product, how influencers ultimately influence the decisions of the latter and most importantly, what the contribution of the factor of trust (or not) between users. We also highlight relevant metrics that have been recorded so far in the literature as well as SR system architectures.