The Research On The Method Of QAR Data Organization Based On Data Warehouse And The Similarity Measurement Of Clustering Pattern

With the occurrence of accident in aviation safety, the safety situation of civil aviation is not optimistic. In order to improve the reliability of scientific management in aviation safety and reduce the accident rate, the relevant flight data are collected and analyzed to identify potentially unsafe factors. It is to determine flight safety problems and provide the basis for the establishment of relevant preventive measures.Combined with QAR data's characteristic of complex format, large volumes of data and strong professional trait, the associated flight safety analysis has some difficulty. This paper fully uses the advantage of data warehouse and mining technology in order to get data with analytical value from massive information. On the basis of systematic succession of fellow students' research results, this paper researches the method of QAR data organization based on data warehouse. It is designed and implemented for QAR data warehouse on the subject of flight phase. To research the similarity of multi-dimensional complex flight data, it proposes a novel method to measure the deviation based on clustering pattern. This method can compare the similarities and differences between multi-dimensional complex flight data sets at a higher abstract level, extracting the representative data sets for analysis, to reduce the repetitive processing of data analysis. Quite different data sets can be extracted for further analysis and found the hidden factor of security risk.In this paper, we do the following tasks: First, to study the relationship between the QAR data distribution and the flight accident rules, it is designed and implemented for QAR data warehouse on the subject of flight phase, to provide the analytical basis of comparing different phases; Second, a framework model is proposed to measure the deviation based on clustering pattern, measuring the difference between two different data sets in the same model. And it applies this method to the similarity analysis of multi-dimensional complex flight data sets, comparing the similarities and differences at a higher abstract level; Third, data reduction technology is used to deal with the large volumes of data to detect the consistency of the results before and after dimensional reduction. The experimental results present that the result after dimensional reduction is better.