Entity Recognition in Chinese Classics via Fine-tuned Large Language Models

Due to the significant differences between the grammatical structures of Chinese classics and modern Chinese, as well as limitations in corpus annotation, research on knowledge mining and knowledge services for classical texts remains limited. Entity recognition in Chinese classics plays a vital role in revealing historical development patterns and constructing historical ontologies. In this study, based on the Zuo Zhuan corpus, we fine-tuned four large language models-Qwen, DeepSeek, Xunzi, and Baichuan and proposed a method for entity recognition in Chinese classics utilizing fine-tuned large language models. Experimental results demonstrate that these fine-tuned large language models can effectively perform entity recognition tasks.