基于SVSM的装备故障案例相似度匹配算法  

Similarity Matching Algorithm of Equipment Fault Case Based on SVSM

作　　者：邓兴宇 胡双演[1] 李钊[2] 隋中山 孙登会[4] 

机构地区：[1]第二炮兵工程大学 [2]二炮驻石家庄地区军事代表室 [3]第二炮兵工程大学士官学院 [4]中国人民解放军96161部队

出　　处：《无线电工程》2016年第2期31-35,共5页Radio Engineering

摘　　要：分析和总结了电子防护装备故障案例的特点,设计了电子防护装备故障诊断的领域本体。利用本体在表示案例时所体现的语义关联信息,提出了基于领域本体的语义特征向量空间模型(Semantic Vector Space Model,SVSM),弥补了传统的向量空间模型在描述故障特征项语义关联度时忽略了文档词条的位置和结构因素的缺陷。引入自定义权值的语义特征项标志位,提出了一种改进的基于案例推理的电子防护装备故障诊断算法,提高了SVSM各维度值的准确性和相似度匹配算法的自由度。通过实验,验证了算法的有效性和可靠性。The domain ontology of electronic protective equipment fault diagnosis is designed by analyzing and summarizing the characteristics of equipment fault cases. Considering that the traditional vector space model（ VSM） ignores the role of the document entry position and structure factors in describing the semantic correlation of fault features,the semantic vector space model（ SVSM）based on semantic characteristics of domain ontology is proposed by exploiting the advantages of ontology in semantic knowledge representation. The accuracy of each dimension value of SVSM and the flexibility of the similarity matching algorithm are improved by using the flags of semantic characteristics weights,and the electronic protective equipment fault diagnosis algorithm based on case-based reasoning（ CBR） is proposed. The validity and reliability of the algorithm are proved through experiment.

关 键 词：案例推理 故障诊断 领域本体 语义向量空间模型 相似度匹配 电子防护系统 

分 类 号：TP391.3[自动化与计算机技术—计算机应用技术]