Research on Fault Diagnosis Technology for Unstructured Data Oriented Avionics Systems

In response to the fault diagnosis requirements of unstructured text data in avionics systems, this article establishes a text classification model through text preprocessing, feature selection, and text representation, and classifies the text test set to achieve fault diagnosis of unstructured text data in avionics systems. This article studies the classification and diagnosis of avionics system fault description texts based on classification algorithms such as naive Bayesian classification, K-nearest neighbor classification, SVM classification, LDA model, etc. This article collected unstructured text data from 1500 avionics system fault descriptions, and validated and analyzed the text classification fault diagnosis algorithm proposed in this article. The fault diagnosis efficiency was significantly improved, and the diagnostic accuracy was high. This article proposes a classification method for unstructured data in avionics systems, which can provide reference for maintenance personnel, provide a basis for management decision-making of avionics maintenance organizations, and improve the maintenance quality of avionics systems.