垃圾分类知识图谱构建研究与实现  

RESEARCH AND IMPLEMENTATION OF GARBAGE CLASSIFICATION KNOWLEDGE GRAPH

作　　者：常晓莹 宋绍京[1] 王华[2] Chang Xiaoying;Song Shaojing;Wang Hua(Shanghai Polytechnic University,Shanghai 201209,China;Shanghai Institute of Computing Technology,Shanghai 200040,China)

机构地区：[1]上海第二工业大学,上海201209 [2]上海市计算技术研究所,上海200040

出　　处：《计算机应用与软件》2022年第5期247-252,共6页Computer Applications and Software

基　　金：上海第二工业大学校重点学科建设项目(XXKZD1605);上海第二工业大学研究生基金(EGD19YJ0102)。

摘　　要：针对当下垃圾分类的热门话题,从无到有搭建一个以垃圾分类为中心的一定规模的垃圾分类知识图谱。项目采用网络爬虫、BiLSTM-CRF、BiGRU-Attention等技术对垃圾分类中文文本数据进行爬虫获取、实体识别、关系抽取,形成“实体-关系-实体”的三元组信息,使用Neo4j可视化展示三元组关系,形成结构化的知识图谱。构建垃圾分类知识图谱,为垃圾分类知识问答系统奠定了基础,使人们在日常生活中能够更好地将垃圾分好类、分对类。In response to the current hot topic of garbage classification,a knowledge graph of garbage classification with a certain size centered on garbage classification has been built from scratch.The project used web crawlers,BiLSTM-CRF,BiGRU-Attention and other technologies to perform crawler acquisition,entity recognition,and relationship extraction on garbage classification Chinese text data to form"entity-relationship-entity"triples.It used Neo4 j to visualize triplet relationship,forming a structured knowledge graph.Construction of a garbage classification knowledge map laid the foundation for a garbage classification knowledge question answering system,so that people could better classify garbage into two categories in daily life.

关 键 词：垃圾分类 网络爬虫 实体识别 关系抽取 Neo4j 

分 类 号：TP3[自动化与计算机技术—计算机科学与技术]