多源异构土壤污染数据知识图谱构建  

Construction of knowledge graph from multi-source heterogeneous soil pollution data

作　　者：霍峥[1] 武彦斌 田俊平[3] 周亚鹏 HUO Zheng;WU Yanbin;TIAN Junping;ZHOU Yapeng(Information Technology School,Hebei University of Economics and Business,Shijiazhuang 050061,China;Management Science and Engineering School,Hebei University of Economics and Business,Shijiazhuang 050061,China;GIS Big Data Platform for Socio-Economy in Hebei,Hebei University of Economics and Business,Shijiazhuang 050061,China;College of Resources and Environment Science,Hebei Agricultural University,Baoding 071001,China)

机构地区：[1]河北经贸大学信息技术学院,河北石家庄050061 [2]河北经贸大学管理科学与工程学院,河北石家庄050061 [3]河北经贸大学经济地理信息大数据平台,河北石家庄050061 [4]河北农业大学国土资源学院,河北保定071001

出　　处：《河北农业大学学报》2021年第2期136-141,共6页Journal of Hebei Agricultural University

基　　金：国家社会科学基金资助项目(18BGL272)。

摘　　要：土壤污染数据来源多样且结构各异,知识图谱是对多源异构土壤污染数据进行数据融合及综合应用的有效手段。针对土壤污染数据的结构特征,本课题组提出了1种土壤污染数据融合与知识图谱构建的方法,首先,基于D-S证据理论和克里格插值法对原始数据进行清洗融合;然后,根据领域本体进行实体识别;最后,抽取土壤污染数据的实体联系与实体属性,形成土壤污染数据的大规模知识图谱。本课题组首次实现了土壤污染数据领域的知识图谱构建,该研究有助于实现土壤污染信息的知识化,提高智能化服务水平。Knowledge graph is an effective method for data fusion and comprehensive application of multi-source heterogeneous soil environmental quality data. In this research, we propose a method of data fusion and knowledge graph construction of soil pollution. Firstly, data cleaning was carried out according to the characteristics of soil pollution data;secondly, the domain ontology of soil pollution data was constructed and entity recognition was carried out according to the domain ontology;finally, the entity relation and entity attribute of soil pollution data were extracted to form large scale knowledge graph of soil pollution data. For the first time, we realize the construction of knowledge graph in the field of soil pollution data. This research will help to realize the knowledge network of soil environmental quality information and improve the intelligent service level.

关 键 词：知识图谱 实体链接 土壤污染 

分 类 号：TG333[金属学及工艺—金属压力加工]