基于BERT-BiGRU-CRF的农业病虫害命名实体识别模型研究

打开文本图片集

中图分类号：TP391.1 文献标识码：A文章编号：2096-4706（2025）22-0007-05

Research on Named Entity Recognition Model of Agricultural Pests and Diseases Based onBERT-BiGRU-CRF

YANGNing，PANJiao，LIANGJiao，RANTao，LENGZhenbei （Chongqing College of International Business and Economics，Chongqing 401520，China）

Abstract：Theaccurate extractionofagricultural pest and disease information holds significant importance foragricultural production.Namedentityrecognition，asakeytechnology，facilitatesthepreciseretrievalofrelevant informationfromvast volumes ofagricultural texts.Addresing limitationsofconventional methods—suchasrelianceonmanual dictionariesand inadequate feature extraction capabilities—this paper proposes a BERT-BiGRU-CRF-based model for agricultural pest and disease namedentityrecognition.ThemodelfirstemploystheBERTpre-trainedmodeltogenerate high-qualitycontextual semantic vectorrepresentations.Subsequently，BiGRU captures long-range contextual dependencies tocomprehensively extract sequence features.Finally，CRFselects theoptimal annotationsequence to produce accurate entityrecognition results.Experimentalesultsdmostratetatteproposedmodelaievessuperiorprecision，ecallandF1soresoaselfconstructed agriculturalpestand disease named entityrecognitiondataset.Optimal anotation sequences toproduceacurate entityrecognitionresults.Experimentalresultsdemonstratethatonaself-constructedagriculturalpestanddisease namedentity recognition dataset， the proposed model achieves precision，recall，and F1 scores of 76.46% 79.65% ，and 77.92% respectively.

Keywords： agricultural pests and diseases; named entity recognition; BERT; BiGRU; CRF

0 引言

农业作为国民经济的基础产业，病虫害的有效防控直接关系到粮食安全与农民增收。（剩余6698字）