Knowledge Graph-Enhanced Zero-Shot Topic Classification: A Multi-Strategy Comparative Study

In three linesComparative study of zero-shot multi-label topic classification using knowledge graphs extracted from documents. Framework tested on 15 LLMs and 8 datasets: keyword-enhanced variant outperforms baseline, graph augmentation helps small models but hurts large ones, and self-consistency decoding increases costs fivefold without performance gains.

Summary generated by Claude — human-verified