GNN-based Approach for User Retweet Behavior Prediction

As Neural Networks have achieved excellent results in many fields, researchers attempt to use Neural Networks to solve graph processing tasks such as node classification, link prediction, and graph classification. Graph Neural Network (GNN) is a machine-learning model that takes graphs as input. In this article, we select active users on a Twitter page and predict their retweet behaviors using Graph Attention Network (GAT). Negative sampling is also applied due to the imbalanced dataset. Experimental results show that GAT well predicts the retweet behavior and tweet count.