基于支持向量机和余弦相似度的故障诊断方法

Fault Diagnosis Methods Based on Support Vector Machine and Cosine Similarity

故障诊断是一种广泛应用于企业的工程技术,有效的故障诊断可以为企业节省大量的人力和物力的开销。传统的文本故障诊断大多采用余弦相似度算法,当匹配出错、数据靠后以及数据量较大时,往往无法满足客户的实时需求。因此,本文采用支持向量机算法对用户输入的故障描述文本语句进行粗划分,筛选出具有相似特征的大类。在此基础上,依据粗分类结果,进一步使用余弦相似度算法进行精确匹配,从而选取匹配相似度最高的故障产生原因和防治措施以反馈客户。实验结果表明,本文所提的故障诊断算法可以有效地进行故障诊断,为企业带来可观的经济效益。

Fault diagnosis is a kind of engineering technology widely used in enterprises. Effective fault diag-nosis can save a lot of expenses in manpower and material resources for the enterprise. Tradition-al text fault diagnosis mostly uses the cosine similarity algorithm. When the matching is wrong, the data falls behind, and the amount of data is large, it often fails to meet the real-time needs of cus-tomers. Therefore, this paper uses the support vector machine algorithm to coarsely divide the fault description text sentences input by the user to screen out the large categories with similar characteristics. Based on the rough classification results, this paper further uses the cosine similar-ity algorithm to perform accurate matching, so as to select the cause of the fault with the highest matching similarity and preventive measures to feedback customers. Experimental results show that the fault diagnosis algorithm proposed in this paper can effectively perform fault diagnosis and bring considerable economic benefits to the enterprise.