Ontological approach to predicting psychological characteristics of social network users based on knowledge graphs

This paper presents an ontological approach to predicting the psychological characteristics of social network users based on knowledge graphs. The developed ontology integrates data from various sources, such as VKontakte user profiles, their activities, psychological test results, and sociological research. Using knowledge graphs and machine learning methods allows for high-accuracy predictions of users' personality traits. The conducted research confirmed the high efficiency of the proposed methodology: prediction accuracy was 78%, recall was 75%, and F-measure was 77%. The use of knowledge graphs allowed not only structuring and interpreting the data but also ensuring the flexibility, scalability, and interpretability of the model. This is especially important for dynamic environments, such as social networks, where data is constantly updated. The obtained results open new opportunities for the application of knowledge graphs in cyberpsychology and marketing.