The approach to building a graph knowledge base using social media data

Abstract

The aim of the work was to develop a model of a knowledge base of an information system that collects information from various social networks. The model should improve search efficiency and facilitate the unification of data from heterogeneous sources. The work presents an ontological model for the unification of data profiles of different social networks. This model avoids data redundancy by including contextual information in annotations to ontology relations. In addition, an approach to information retrieval using syntagmatic patterns in the formation of a database tree of posts of social network users is proposed. The article also presents the results of experiments with data from the social network Facebook confirming the effectiveness of the proposed models and algorithms.