儿童用品化学安全知识图谱构建与应用分析

摘要：

为防范儿童用品化学危害可能造成的严重后果，有效支持儿童用品源头治理和过程控制，保障儿童用品安全，本文提出了儿童用品化学安全知识图谱构建框架，根据儿童用品化学危害召回数据，儿童用品化学安全标准文本数据特征和结构化程度，结合UIE模型和基于规则的方法对儿童用品召回数据和安全标准进行知识抽取，通过知识融合构建知识图谱。基于构建的知识图谱清晰分析了儿童用品的安全隐患、化学危害的识别和控制、质量安全监管决策支持。儿童用品化学安全知识图谱可以为儿童用品的生产制造和质量监管提供支持，降低儿童接触有害化学物质的风险，保障儿童健康安全。

Abstract:

In order to prevent the possible serious consequences of chemical hazards in children’s products, effectively support the source management and process control of children’s products, and ensure the safety of these products, in the paper, a framework for the construction of chemical safety knowledge graph for children’s products is proposed. According to the characteristics and structuring degree of the recall data of children’s products and the text data of chemical safety standards for children’s products, combined with the UIE model and rule-based methods, the knowledge extraction of children’s product recall data and safety standards is carried out, and then the knowledge graph is constructed through knowledge fusion. Based on the constructed knowledge graph, the potential safety hazards of children’s products, the identification and control of chemical hazards are clearly analyzed to provide the decision support of quality and safety supervision. The chemical safety knowledge graph of children’s products can provide support for the production and quality supervision of these products, reduce the risk of children’s exposure to harmful chemicals, and ensure children’s health and safety.

基金项目：本文系国家重点研发计划“儿童用品中吸入性化学应激源限量研究”（项目编号：2022YFF0606205）和中央基本科研业务经费项目“基于数据驱动的产品质量安全风险分类技术研究与应用”（项目编号：552023Y-10371）研究成果之一。

作者简介：冀晓东，工程师，研究方向为标准认证与质量管理。孙高岭，研究方向为标准认证与质量管理。涂新雨，博士研究生，研究方向为质量管理、文本挖掘、知识图谱。郑怀城，硕士研究生，研究方向为质量管理、文本挖掘、知识图谱。许应成，研究员，研究方向为质量安全管理，大数据分析等。