Sequence Labeling for Dependency Graph Parsing in Sentiment Analysis

This method achieves fine-grained sentiment graph construction by casting structured sentiment analysis as a sequence labeling task using linearized graph encodings. The approach performs competitively across seven datasets and five languages compared to more complex single-model parsing architectures.