Article Details

Abstract

In order to enhance the comprehensive performance of structured information extraction from resumes under small training sample conditions, and provide a viable approach for efficient resume screening, this paper proposes a method to generate high-quality training corpora in the absence of domain-specific annotation data by constructing a competition-specific domain dictionary and leveraging it to augment datasets from a limited number of resumes. Based on the BERT model, key information extraction is performed after training on datasets of varying augmentation scales, and evaluation metrics are calculated. The experimental results indicate that the F1 scores for information extraction on two test sets reach 91.79% and 76.47%, validating the effectiveness of the “domain dictionary + data augmentation” strategy in low-resource named entity recognition tasks.

Keywords

Named Entity Recognition; Natural Language Processing; domain dictionary; data augmentation; BERT model; information extraction

References

1. [1] Zhou Yeheng, Shi Jiahan, Xu Ruifeng Text matching method combining pre trained model and language knowledge base [J]. Chinese Journal of Information Science, 2020, 34 (2): 63-72
2. [2] Chen Shudong, Ouyang Xiaoye Overview of Named Entity Recognition Technology [J]. Radio Communication Technology, 2020, 46 (3): 251-260
3. [3] Wang Yingjie, Zhang Chengye, Bai Fengbo, etc A Review of Chinese Named Entity Recognition Research [J]. Computer Science and Exploration, 2023, 17 (2): 324-341
4. [4] Devlin J，Chang M W，Lee K，et al. BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding [PP/OL].（2019-05-24）https://arxiv.org/abs/1810.04805.
5. [5] Lee J，Yoon W，Kim S，et al. BioBERT: A Pre-trained Biomedical Language Representation Model for Biomedical Text Mining [J].Bioinformatics，2020，36（4）：1234-1240.
6. [6] Zhang T，Xia C Y，Yu P S，et al. PDALN: Progressive Domain Adaptation over a Pre-trained Model for Low-Resource Cross-Domain Named Entity Recognition [C]//Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing，2021：5441-5451.
7. [7] Karamitsos I，Roufas N，Al-Hussaeni K，et al. LegNER: A Domain-adapted Transformer for Legal Named Entity Recognition and Text Anonymization [J/OL].Frontiers in Artificial Intelligence，2025，8：1638971.https://doi.org/10.3389/frai.2025.1638971.
8. [8] Liu Z H，Jiang F J，Hu Y X，et al. NER-BERT: A Pre-trained Model for Low-Resource Entity Tagging [PP/OL].2021-12-01.https://arxiv.org/abs/2112.00405.
9. [9] Parhati Turajiang, Sun Yuanyuan, Cai Aichen, etc Low resource legal document named entity recognition with domain vocabulary expansion [J]. Computer Engineering and Applications, 2026, 62 (1): 264-273
10. [10] Liang C，Yu Y，Jiang H M，et al. BOND: BERT-Assisted Open-domain Named Entity Recognition with Distant Supervision [PP/OL].（2020-06-28）.https://arxiv.org/abs/2006.15509.
11. [11] Hiramatsu M，Wakabayashi K，Harashima J. Named Entity Recognition by Character-based Word Classification Using a Domain Specific Dictionary [J/OL].EasyChair Preprint，2019：1-12.https://easychair.org/publications/preprint/3Jfw.
12. [12] Dang X C，Wang L，Dong X H，et al. Improving Low-resource Chinese Named Entity Recognition Using Bidirectional Encoder Representation from Transformers and Lexicon Adapter [J/OL].Applied Sciences，2023，13（19）：10759.https://doi.org/10.3390/app131910759.
13. [13] Vaswani A，Shazeer N，Parmar N，et al. Attention is All You Need [C]//Proceedings of the 31st International Conference on Neural Information Processing Systems，2017：6000-6010.https://dl.acm.org/doi/10.5555/3295222.3295349.
14. [14] Li Jun, Lv Xueqiang A Keyword Extraction Method Integrating BERT Semantic Weighting and Network Graph [J]. Computer Engineering, 2020, 46 (9): 89-94
15. [15] Raffel C，Shazeer N，Roberts A，et al. Exploring the Limits of Transfer Learning with a Unified Text-to-text Transformer [J/OL].The Journal of Machine Learning Research，2020，21（140）：1-67.https://www.jmlr.org/papers/v21/20-074.html.
16. [16] Roberts A，Raffel C，Shazeer N. How Much Knowledge can You Pack into the Parameters of a Language Model? [C]//Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing (EMNLP)，2020：5418-5426.

About the authors

Qian Huayuan (2004-), male, Han ethnicity, from Shijiazhuang, Hebei Province, is currently pursuing a bachelor's degree with a research focus on natural language processing.