Аналіз методологічних підходів прогнозування вірусності контенту за допомогою BIG DATA

Abstract

A comprehensive analysis of scientific research on predicting the virality of online content using big data analysis methods has been conducted in the article. The evolution of methodological approaches is examined: from classical time series-based models to modern deep learning architectures, including graph neural networks, transformers, and multimodal models. Special attention is paid to comparative analysis of methods, problems of early prediction, cross-platform content dissemination, model interpretability, and overcoming prediction drift.