Özet:

Air transportation is a very preferred type of transportation for long-distance trips. This type of transportation has made great progress, especially in the last 20 years with the development of technology. Thanks to its fast and safe, passenger capacity is gradually increasing. Despite this situation, the mortality rate is quite high in the case of an aircraft accident. For this reason, hundreds of people can die in a single accident. In this study, aircraft accidents that occurred in the last 20 years in the world were examined. The data including the number of accidents, the number of deaths and the process of the flight where the accidents occurred were used. These data were analyzed using data mining algorithms such as multi-layer perceptron, k nearest neighborhood, Naive Bayes, J48 and regression methods. Accordingly, it was determined that the algorithm that gives the best results for error scale and performance analysis among five different algorithms is J48. Using this algorithm, the occurrence flight phase of aircraft accidents is classified in more detail. Thanks to this study, it has been revealed that choosing the J48 algorithm for the classification of similar data sets will give better results. Also, this study provides significant benefits in terms of getting to the center of the problems, as the stages of accidents are more detailed. Accordingly, it is possible to reduce accidents if policy makers carry out studies taking into account the stages in which accidents occur.

Anahtar Kelimeler: