节点文献

基于知识图谱的农作物良种问答系统的设计与实现

Design and implementation of a recommendation system for improved crop varieties based on Knowledge Graph

  • 推荐 CAJ下载
  • PDF下载
  • 不支持迅雷等下载工具,请取消加速工具后下载。

【作者】 李成林赵珍威李国厚侯志松

【Author】 LI Chenglin;ZHAO Zhenwei;Li Guohou;HOU Zhisong;School of Information Engineering,Henan University of Science and Techtiology;

【通讯作者】 侯志松;

【机构】 河南科技学院信息工程学院

【摘要】 种子是农业生产的基础,通过科学评估农作物品种的适应性、产量潜力、品质特性、抗性等因素,选择适宜当地生态环境和市场需求的优良品种,对提高农业生产效益、促进农村经济发展和农民增收具有重要意义.因此,基于知识图谱技术,构建一个农作物良种问答系统,以帮助农户快速、准确地获取农作物品种相关信息,提高农业生产效率,增加农民收益.为此,研究收集农作物品种相关数据,包括品种名称、品种来源、生长周期、适宜种植区域、发病阶段等信息;采用数据预处理、知识融合和知识存储等方法构建农作物品种知识图谱;借助自然语言处理和BiLSTM-CRF技术设计并开发了一个问答系统.经测试,构建的农作物良种问答系统准确率可达87.67%,能满足用户对农作物品种信息的查询、获取和推荐需求.

【Abstract】 Seeds are the foundation of agricultural production.By scientifically evaluating factors such as crop variety adaptability,yield potential,quality characteristics,and resistance,selecting excellent varieties suitable for local ecological environments and market demands is of great significance for improving agricultural production efficiency,promoting rural economic development,and increasing farmers’ income.Therefore,based on knowledge graph technology,this paper constructs a crop variety QA system to help farmers quickly and accurately obtain relevant information on crop varietie.First,this paper collects relevant data on crop varieties,including variety names,sources,growth cycles,suitable planting regions,and disease stages.Secondly,knowledge graph of crop varieties is constructed using data preprocessing,knowledge fusion,and knowledge storage methods.Then,a QA system is designed and developed using natural language processing and BiLSTM-CRF technology.Finally,the crop variety QA system constructed in this paper has an accuracy rate of 87.67% according to testing,and can meet users’ requirements for querying,obtaining,and recommending information on crop varieties.

【基金】 河南省科技攻关项目(232102111125)
  • 【文献出处】 河南科技学院学报(自然科学版) ,Journal of Henan Institute of Science and Technology(Natural Science Edition) , 编辑部邮箱 ,2024年01期
  • 【分类号】S126;TP391.1
  • 【下载频次】227
节点文献中: 

本文链接的文献网络图示:

本文的引文网络