The analysis of fatal aviation accidents more than 100 dead passengers: an application of machine learning

Abstract

Safety is the most prominent factor that affected accidents in civil aviation history. In safety concept, the primary factors are defined as human, technical, and sabotage/terrorism factors. Despite these primary causes, there have other factors that have an impact to accidents. The study aims to determine the affected factors of the 220 accidents that were ended with more than 100 dead passengers by the primary causes and the other factors such as aircraft type, total distance, the phase of flight, number of total passengers, and time period of the accident. All these factors aims to classify the rate of survivor/non-survivor passenger rate according to most fatal accidents. It is used logistic regression and discriminant analysis for multivariate statistical analyses comparing the machine learning approaches to show the algorithms' robustness. At the end of the analysis, it is seen that machine learning techniques have better performance than multivariate statistical methods in related to accuracy, false-positive rate, and false-negative rates. The managerial aim of this study is related to find the most important factors that affected the most fatal accidents. These factors are found as; the phase of flight, the primary cause, and total passenger numbers according to machine learning and multivariate statistical models for classifying the rate of survivor/non-survivor passenger numbers.