知识图谱下非结构化物联网数据关联规则挖掘  

Mining of Unstructured Internet of Things Data Association Rules Under the Knowledge Graph

作　　者：田鹏 TIAN Peng(Guizhou Postal and Telecommunications Planning and Design Institute Co.,Ltd.,Guiyang 550003,China)

机构地区：[1]贵州省邮电规划设计院有限公司,贵阳550003

出　　处：《智能物联技术》2025年第1期129-132,共4页Technology of Io T& AI

摘　　要：非结构化物联网数据在处理过程中易受到不一致性等因素的影响,使得数据间关联规则的挖掘结果不准确。为此提出知识图谱下非结构化物联网数据关联规则挖掘。构建数据分布结构模型,结合数据文本特征量实现对不同数据源的融合集成。利用离散向量化方式构建非结构化物联网数据知识图谱,通过实体消歧生成候选实体,识别出数据实体。获取频繁项降序数据集序列,筛选出高价值关联规则,实现数据关联规则的挖掘。实验结果表明,所研究方法的Rand指标值高于0.86,排名倒数平均值(Mean Reciprocal Rank,MRR)指标值均高于0.82,数据关联规则挖掘结果较为准确。Unstructured networked data are easily affected by inconsistency and other factors during processing,which makes the mining results of association rules between data inaccurate.Therefore,the mining of association rules of unstructured networked data based on knowledge map is proposed.The data distribution structure model is constructed,and the integration of different data sources is realized by combining the data text features.The knowledge map of unstructured networked data is constructed by discrete quantization,and candidate entities are generated by entity disambiguation to identify data entities.Get descending data set sequence of frequent items,screen out high-value association rules,and realize data association rules mining.The experimental results show that the Rand index value of the research method is higher than O0.86,and the Mean Reciprocal Rank(MRR)index value is higher than 0.82,so the data association rules mining results are more accurate.

关 键 词：知识图谱 非结构化数据 物联网数据 关联规则 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]