Simple Named Entity Recognition (NER) System with RoBERTa for Ancient Chinese
Yunmeng Zhang, Meiling Liu, Hanqi Tang, Shanshan Lu, Lang Xue · 2025
Named Entity Recognition (NER) is a fundamental task in Natural Language Processing (NLP), particularly in the analysis of Chinese historical texts.In this work, we propose an innovative NER model based on Gu-jiRoBERTa, incorporating Conditional Random Fields (CRF) and Long Short Term Memory Network(LSTM) to enhance sequence labeling performance.Our model is evaluated on three datasets from the EvaHan2025 competition, demonstrating superior performance over the baseline model, SikuRoBERTa-BiLSTM-CRF.The proposed approach effectively captures contextual dependencies and improves entity boundary recognition.Experimental results show that our method achieves consistent improvements across almost all evaluation metrics, highlighting its robustness and effectiveness in handling ancient Chinese texts.