面向食品安全法规的知识图谱构建

Construction of Knowledge Graph for Food Safety Regulations

我国建立了完备的食品安全法规体系,其具有海量和零散性的特点,难以检索分析。以食品安全法规文本数据为依托,通过自顶向下和自下而上的方式进行食品安全法规知识图谱的构造研究。首先,获取多源异构的食品安全法律法规和问答数据语料,对用户的需求进行分析。其次,定义食品安全知识图谱的本体层及其属性,使用基于规则的方法对知识进行抽取,针对规则性不强的知识,使用基于机器学习的命名实体识别方法完成领域命名实体识别。最后,实现食品安全法规知识图谱的构建。

China has established a complete food safety regulatory system,which is characterized by massive and fragmented nature and difficult to retrieve and analyze.Therefore,we conduct research on the construction of a Knowledge Graph of food safety regulations through a top-down and bottom-up approach based on the textual data of food safety regulations.First,we obtain multi-source heterogeneous food safety laws and regulations and Q&A data corpus,and analyzes user needs.Then,we define the ontology layer and attributes of the food safety Knowledge Graph.We extract the knowledge by using a rule-based method,and complete the domain named entity recognition by using Machine Learning-based methods for the knowledge with weak regularity.Finally,we realize the construction of Knowledge Graph for the food safety regulations.