Intelligent fault diagnosis method for aviation maintenance text record

The invention discloses an intelligent fault diagnosis method for aviation maintenance text records. The method comprises the following steps: constructing an aviation maintenance text data fault database and a fault cause database; preprocessing the fault text; dividing a fault data training set and a test set; performing aviation maintenance record input vector representation; and aeronautical maintenance record fault diagnosis. According to the method, a Transform encoder and a DAL model in an ALBERT model are fused to perform feature extraction on maintenance records, probability distribution of all fault types is obtained, and thus a fault diagnosis result is obtained. According to the method, an aviation field dictionary does not need to be constructed, vector representation is carried out by taking words as units, the model is improved in accuracy while the diagnosis speed is ensured by adopting a model fusion mode, end-to-end intelligent diagnosis of aviation maintenance text data is realized, and the