Research On Association Analysis And Its Application In Aviation Safety Reports

Civil aviation safety is the core competitiveness of enterprises, which bearing on the development and expansion of civil aviation enterprises. In order to enhance civil aviation system, detect and correct system deficiencies quickly and accurately, control and eliminate hidden dangers to aviation safety, countries around the world set up aviation safety reporting system to collect reports of unsafe incidents and potential safety problems.Existing aviation safety report contains an analysis, the main structure of which is simple data queries and statistics. The analysis of unstructured data require experts spending a lot of time, the rule can be found, but it can not reveal the intrinsically linkage and the latent model of data. Artificial analysis is time-consuming and labor-intensive, the accuracy of analysis depend on staff capacity and experience. Therefore, the study of effective analysis methods, discovering the unknown data and potential problem automatically and accurately, revealing the events related to the airline control and eliminating potential safety problems to improve aviation safety performance has an extremely important practical significance.Text-based association analysis technology was applied to aviation safety report mining, and multilayer association rule mining based on ontology was proposed. The contents are focused on the following aspects:1,In view of the identification of unknown word in report preprocessing, the method to create special domain lexicon is proposed, and the aviation safety dictionary was built.2,Association analysis technology based on terms was applied to aviation safety report mining to access frequent term modes.3,The ontology-based multilayer association rule mining was proposed.4,Visualization technology was used to display the results of association analysis.5,The prototype system of aviation safety analysis-ASARM was realized.