基于指标体系的相似故障问题推荐方法研究

Research on recommendation method for similar fault problems based on indicator system

针对试飞故障问题数量大、内容多、无法高效利用其信息价值的问题,提出基于运行阶段、ATA章节号、故障文本和CAS、OMS信息的指标体系相似故障问题推荐方法。依据各指标特征,设计不同的相似度计算方法。此外,利用正则表达式,提出一种自动从故障文本描述中提取CAS信息和OMS信息的方法。最后通过实例验证相对于只选取运行阶段、ATA章节号和故障文本作为指标的推荐算法以及只使用TF-IDF方法的算法,本文所提算法在准确率上分别提高了25%和28%,在召回率上分别提高了约27%和31%,可以为相似试飞故障问题的推荐提供参考。

In response to the problem of a large number of test flight fault problems and the inability to efficiently utilize their information value,a recommendation method for similar fault problems based on the index system of flight stages,ATA,fault text,CAS and OMS information was proposed.Similarity calculation methods were proposed for each indicator based on their characteristics.In addition,a method is proposed to automatically extract CAS and OMS information from fault text descriptions using regular expressions.Finally,an example was used to validate the recommendation algorithm that only selects flight stage,ATA,and faults text as indicators,as well as the algorithm that only uses TF-IDF method.The algorithm proposed in this article has improved 25%and 28%in accuracy,27%and 31%in recall.It can provide reference for the recommendation of similar flight test fault problems.