[2404.13079] Relational Graph Convolutional Networks for Sentiment Analysis

Relational Graph Convolutional Networks (RGCNs) are proposed for sentiment analysis, offering interpretability and flexibility in capturing dependencies between data points represented as nodes in a graph. The effectiveness of RGCNs is demonstrated using pre-trained language models such as BERT and RoBERTa on product reviews from Amazon and Digikala datasets.