Key Feature Extraction Improves Aircraft Quality Detection Capabilities

Testing as an important factor in ensuring aircraft quality, means very important for ensuring the quality of aircraft. The original manual inspection method required a complete reading of the inspection requirements document, which resulted in low detection efficiency and accuracy. In order to improve detection efficiency, this article proposes to extract key features from detection/inspection requirements documents from the perspective of natural language processing to form a structured database. This article introduces the process of extracting key features and the algorithms used. After manually annotating only about 100 pieces of data, the accuracy of the model reached $67 \%$, which greatly accelerated the inspection efficiency compared with the old way.