Abstract:

Each social network used by users for different purposes contains different user data. The finding of users’ accounts on different social networks and the combination of the found data and the aggregation in a single data pool will be a very important factor that will both improve the recommending systems and throw out the user experience. In the framework of the study, original node equalization and node similarity methods were suggested. In the topological-based node recommendation, the diameter method was used while the intensity relationships between connections were also taken into account. The similarity-based node similarity method has increased the number of successful node matchings with the specificity selection criteria, the problem of starting point detection and variable formulation. However, this study identified equalization and similarity according to both the user’s profile characteristics and the relationships between them and other users. There are nine different methods for finding the same accounts on different social networks. The recommended methods were tested on social networks collected up to two to six social networks data and the user’s matching success rates were measured. In these results, the success rates reached 95%. This makes it possible to create a wide user profile that covers multiple social networks for users gathered on the same graph with different properties.

Keywords: