专利知识图谱构建

只是想记录一下自己的课程作业，如果对读者也有一些帮助就更好啦！

使用的工具为python+neo4j。

1.数据来源

以IncoPat专利数据库为数据来源，通过构建检索式(IPC=(A01C15/00)) AND (PD=[20220101 TO 20230101])筛选出2022年施肥机械相关的数据，共获取到1901条专利数据。

2.本体设计

该本体结构包括5 种实体类型，4种语义关系。

3.实体与关系构建

4.数据存储

完整代码如下：

from py2neo import Graph,Node,Relationship,Subgraph
import csvpatentgraph = Graph("http://localhost:7474", auth=("neo4j", "neo4j"))with open('fertilizermachine.csv') as csvfile:reader = csv.DictReader(csvfile)for row in reader:#create entitiespatentid = Node("patentid", name=row["公开（公告）号"])patentid['title'] = row["标题 (中文)"]#create attributespatentgraph.create(patentid)claimer = patentgraph.nodes.match("claimer", name=row["申请人"]).first()if not claimer:claimer = Node("claimer", name=row["申请人"])patentgraph.create(claimer)publicdate = patentgraph.nodes.match("publicdate", name=row["公开（公告）日"]).first()if not publicdate:publicdate = Node("publicdate", name=row["公开（公告）日"])patentgraph.create(publicdate)patenttype = patentgraph.nodes.match("patenttype", name=row["专利类型"]).first()if not patenttype:patenttype = Node("patenttype", name=row["专利类型"])patentgraph.create(patenttype)country = patentgraph.nodes.match("country", name=row["公开国别"]).first()if not country:country = Node("country", name=row["公开国别"])patentgraph.create(country)#create relationshipsclaim = Relationship(claimer, "claim", patentid)dateis = Relationship(patentid, "dateis", publicdate)typeis = Relationship(patentid, "typeis", patenttype)countryis = Relationship(patentid, "countryis", country)A = Subgraph(relationships=[claim, dateis, typeis, countryis])patentgraph.create(A)print(row)

5.图谱展示

 done√

本文来自互联网用户投稿，文章观点仅代表作者本人，不代表本站立场，不承担相关法律责任。如若转载，请注明出处。 如若内容造成侵权/违法违规/事实不符，请点击【内容举报】进行投诉反馈！