Talk · International Russian Smart Industry Conference · March 24, 2025
Prediction of Online Social Network's Users Features and Social Behavior by Graph Neural Networks
Fail Mubarakovich Gafarov, Daniyar Mingazov, Pavel Nikolaevich Ustin
Abstract
In this work we developed and trained a machine learning models based on graph neural networks to predict some features as gender, age group, social behavior of VKontakte online social network users. As the input data for these models we used users' characteristics like different metrics, which represent quantitative data, and texts posted by user on his personal page, which represent a qualitative data. That is, in this work we solved the graph nodes' property prediction task. The performance of different graph neural network architectures based on GCN, SAGE, GAT, Transformer and GIN had been compared. The maximal performance had been obtained for Transformer based model, F1-score achieved 0.7809 in gender prediction, 0.7505 in age group prediction and 0.8008 in social behavior level prediction on test dataset.
More by Fail Gafarov
- Predicting Changes in Personality Behavior Under the Influence of an InterlocutorRussian Digital Libraries Journal · 2026
- RNN-Based Approach to Cognitive Architecture of Subject Behavior in Digital EnvironmentInternational Conference on Computer Science and Engineering · 2026
- Cross-Model Analysis of Token-Level Explanations in Tatar Text ClassificationUral Symposium on Biomedical Engineering, Radioelectronics and Information Technology · 2026
- СРАВНИТЕЛЬНЫЙ АНАЛИЗ МОДЕЛЕЙ АРХИТЕКТУРЫ TRANSFORMER В ЗАДАЧЕ УПРОЩЕНИЯ ТЕКСТАВысшая школа: научные исследования · 2025
