面向航空器材管理的数据规范化算法研究OA

In order to solve the problem of inconsistent naming of species data in aviation equipment management during aircraft manufacturing,a semantic similarity based data normalization algorithm for aviation equipment species is proposed.The algorithm can cluster similar species data to realize the normalization of the aviation equipment species names.The Sentence-BERT(SBERT)model is used to perform the semantic vectorization of equipment names and performance parameters in the aviation equipment species data to capture the deep semantic information of the data.As the clustering directly with the generated high-dimensional vectors may lead to lower clustering accuracy,the t-SNE algorithm is introduced to perform dimensionality reduction on the semantic vectors generated by SBERT,which can effectively preserve the semantic structure of the vectors while reducing dimensions,thereby improving the clustering effect.The hierarchical clustering algorithm is used to cluster the dimensionality-reduced data,achieving the normalization of aviation equipment species data.The comparative experimental results on publicly available datasets show that,in comparison with mainstream traditional methods,the proposed SBERT and hierarchical clustering-based normalization algorithm can realize a maximum clustering accuracy(ACC)of 70.6%,and the highest values of 82.1%for normalized mutual information(NMI)and 58.5%for adjusted rand index(ARI).It proves that the proposed method has significant advantages in clustering performance and semantic information preservation,effectively meeting the needs for normalization of aviation equipment species data and providing strong technical support for aviation equipment data management.