现代信息科技2025，Vol.9Issue(22)：7-11,5.DOI:10.19850/j.cnki.2096-4706.2025.22.002
基于BERT-BiGRU-CRF的农业病虫害命名实体识别模型研究
Research on Named Entity Recognition Model of Agricultural Pests and Diseases Based on BERT-BiGRU-CRF
摘要
Abstract
The accurate extraction of agricultural pest and disease information holds significant importance for agricultural production.Named entity recognition,as a key technology,facilitates the precise retrieval of relevant information from vast volumes of agricultural texts.Addressing limitations of conventional methods—such as reliance on manual dictionaries and inadequate feature extraction capabilities—this paper proposes a BERT-BiGRU-CRF-based model for agricultural pest and disease named entity recognition.The model first employs the BERT pre-trained model to generate high-quality contextual semantic vector representations.Subsequently,BiGRU captures long-range contextual dependencies to comprehensively extract sequence features.Finally,CRF selects the optimal annotation sequence to produce accurate entity recognition results.Experimental results demonstrate that the proposed model achieves superior precision,recall,and F1 scores on a self-constructed agricultural pest and disease named entity recognition dataset.Optimal annotation sequences to produce accurate entity recognition results.Experimental results demonstrate that on a self-constructed agricultural pest and disease named entity recognition dataset,the proposed model achieves precision,recall,and F1 scores of 76.46%,79.65%,and 77.92%respectively.关键词
农业病虫害/命名实体识别/BERT/BiGRU/CRFKey words
agricultural pests and diseases/named entity recognition/BERT/BiGRU/CRF分类
计算机与自动化引用本文复制引用
杨宁,潘娇,梁姣,冉涛,冷震北..基于BERT-BiGRU-CRF的农业病虫害命名实体识别模型研究[J].现代信息科技,2025,9(22):7-11,5.基金项目
重庆对外经贸学院2024-2025年度科学研究项目(KYZK2024041) （KYZK2024041）
