Language Prediction of Twitch Streamers using Graph Convolutional Network
Md. Nahid Hasan, Nakshi Saha, Md. Anisur Rahman · 2025
Social Network is a common example of the graph structure. There are many algorithms available for predicting different elements of social networks. Language is one of the elements. Prediction problems generally stay within 2D information. Graphs provide more structural information like edge relationship edge feature etc. While predicting social network elements, information about graph structure can be very useful. Graph Convolutional Network gives the opportunity to use graph structure as a piece of new information. Twitch is a popular social media platform where people stream gaming-related content. In our work, we have proposed a Graph Convolutional Network model for predicting the language of Twitch gamers. We also showed that it outperforms other machine learning and deep learning algorithms like Decision trees, Random Forest, Support Vector Machine, Neural Network and others. While other models could only reach 45.8% accuracy, GCN was able to achieve 73.43% by utilizing the graph structure.