Named Entity Recognition of Coal Mining Equipment Based on Fine-Tuning of Large Language Models

Named Entity Recognition plays a crucial role in the construction of knowledge graphs by identifying and classifying named entities, such as persons, locations, organizations, and times, from unstructured text. NER serves as the foundation for subsequent tasks such as relation extraction and knowledge linking. In recent years, with advancements in natural language processing and deep learning technologies, NER methods, such as deep learning-based sequence labeling models have made significant progress. However, traditional NER methods rely on large amounts of annotated data for training, which is time-consuming, laborintensive, and prone to quality and rule limitations. In the field of coal mining comprehensive mining equipment, which is critical for efficient and safe coal extraction, there is <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$a$</tex> lack of high-quality annotated data, making model training challenging. To address this issue, this chapter proposes <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$a$</tex> NER method based on fine-tuning large language models. By using the pretrained ChatGLM3 model and fine-tuning it with a small amount of domain-specific data related to coal mining equipment, the accuracy of named entity recognition can be significantly improved, reducing the reliance on large annotated datasets and optimizing the knowledge graph construction process.