面向我国畜牧业法律法规的知识图谱构建  

Construction of Knowledge Graph for Animal Husbandry Laws and Regulations in China

作　　者：张容祯 孟小艳[1] 刘潇潇 汪洋 ZHANG Rongzhen;MENG Xiaoyan;LIU Xiaoxiao;WANG Yang(College of Computer and Information Engineering,Xinjiang Agricultural University,Urumqi830052,China)

机构地区：[1]新疆农业大学计算机与信息工程学院,新疆乌鲁木齐830052

出　　处：《畜牧与饲料科学》2023年第3期69-74,共6页Animal Husbandry and Feed Science

基　　金：新疆维吾尔自治区高校基本科研业务费科研项目“农业大数据交换共享与可视化平台构建”(XJEDU2022J009);乌鲁木齐市科学技术计划项目(Y16330001)。

摘　　要：[目的]提出一种对畜牧业法律法规知识组织的方法,使相关知识实现更加有序的组织。[方法]选取国家法律法规数据库、北大法宝、北大法意、威科先行和中国法律知识资源总库中的法律法规资源作为畜牧业法律法规的数据来源,对畜牧业相关法律法规进行筛选、获取,提取对应的制定机关、时效性、效力位阶、法规类别、公布日期、施行日期和年份信息,并对获得的数据进行预处理;对本体进行构建,定义实体和关系;基于规则的方法对获得的法律法规文本抽取实体和关系,并将抽取的实体和关系暂储到MySQL数据库中;利用Python中的SQLAlchemy库和py2neo库分别操作MySQL数据库和Neo4j图数据库,将MySQL中的数据转为实体关系三元组,存储到Neo4j图数据库中完成畜牧业法律法规知识图谱的存储。[结果]共收集到287个畜牧业相关的法律法规,按照法律法规是否含有章进行区分,其中,含有章的有211个,共1470章,10457条,不含章的76个,共2145条;定义了“法律法规”“发布部门”“时效性”“效力级别”“法规类别”“章”“条”等共10种类型的实体,定义了“畜牧业法律法规-法律法规的类别关系”“畜牧业法律法规-法律法规的时效性关系”“畜牧业法律法规-法律法规的发布部门关系”“畜牧业法律法规-法律法规的效力级别关系”等共10种类型的关系;构建了一个含有实体数量14936个、关系数量16339个的畜牧业法律法规知识图谱。[结论]该方法可应用于畜牧业法律法规知识的组织,可使知识的组织更加有序,关联更加紧密。[Objective]The present study was conducted to propose a method for organizing the knowledge of animal husbandry laws and regulations to achieve a more orderly organization of relevant knowledge.[Method]The online legal resources in the websites of flk.npc.gov.cn,pkulaw.com,lawyee.org,wkinfo.com.cn,and law.cnki.net were used as the data sources of animal husbandry laws and regulations,the relevant laws and regulations of animal husbandry were screened and obtained.Subsequently,the corresponding formulation authority,timeliness,effectiveness hierarchy,regulation category,issuing date,implementation date and year information were extracted,and the obtained data was preprocessed.The ontology was built,and the entities and relationships were defined.Using rule-based method,the entities and relationships from the obtained laws and regulations texts were extracted,and then temporarily stored in MySQL database.The SQLAlchemy database and py2neo database in Python were used to operate MySQL database and Neo4j Graph database to convert the data in MySQL into entity relationship triad,and then stored in Neo4j Graph database to complete the storage of knowledge graph of animal husbandry laws and regulations.[Result]A total of 287 laws and regulations related to animal husbandry were collected and classified based on whether the laws and regulations contained chapters.Among these,211 laws and regulations contained chapters,totaling 1470 chapters and 10457 articles,whereas 76 laws and regulations did not contain chapters,totaling 2145 articles.There were 10 types of entities defined,including′laws and regulations′,′issuing authority′,′timeliness′,′effectiveness hierarchy′,′regulation category′,′chapter′,′article′,etc.There were 10 types of relationships defined,including′the relationship of animal husbandry laws and regulations to category of laws and regulations′,′the relationship of animal husbandry laws and regulations to timeliness of laws and regulations′,′the relationship of animal h

关 键 词：畜牧业 知识图谱 法律法规 Neo4j 

分 类 号：S818.9[农业科学—畜牧学] G250.73[农业科学—畜牧兽医] TP391.1[文化科学—图书馆学]