Automatic classification method and device for aviation safety reports

Abstract

The invention provides an automatic classification method and device for aviation safety reports, provides a kg2vec + CNN aviation safety report automatic classification scheme, uses a kg2vec model totrain word vectors of the aviation safety reports to serve as a feature matrix of aviation safety report text data, and inputs the feature matrix into a convolutional neural network. A new text feature vector is obtained after convolution kernel pooling, at the moment, the feature vector not only contains rich semantic features but also contains grammatical features, finally, the feature vector is input into a softmax classifier to be classified, a classification result is obtained, and therefore automatic classification for the aviation safety reports is achieved. Detailed experiments are carried out on a real aviation safety report data set, and experimental results show that kg2vec word vectors can effectively improve the classification accuracy of aviation safety reports; in all control experiments, the F1-score value of the kg2vec + CNN automatic classification scheme is the highest, and the classification accuracy is as high as 91.4%.

Description

technical field [0001] The present application relates to the technical field of aviation safety, in particular to an automatic classification method and device for aviation safety reports. Background technique [0002] In recent years, with the rapid development of the economy and the improvement of living standards, more and more people prefer airplanes with fast speed and convenient travel. As a result, the global civil aviation passenger traffic has increased year by year, which has brought great challenges to aviation safety. The Aviation Safety Reporting System (ASRS) has collected a large number of reports of unsafe incidents and potential safety hazards discovered by civil aviation front-line employees in their daily work. These reports are first-hand valuable information on aviation safety, and provide an important basis for eliminating potential safety hazards and formulating effective corrective measures and macro policies. [0003] The traditional aviation safe...

AI Technical Summary

Problems solved by technology

This traditional classification method is not only costly, but also inefficient. There is also the problem that the same person classifies the same report in different time periods, and the classification results obtained are often inconsistent.

In addition, in the face of the increasing number of aviation safety reports, relying only on manual classification and management of aviation safety reports is far from meeting the current management needs of aviation safety reports

Images