(1138-4082-2-PB)A STUDY ON PREDICTION MODELING OF KOREA MILLITARY AIRCRAFT ACCIDENT OCCURRENCE

摘要

This research reports the analysis on the causes of accidents and case studies during the last 30 years in order to predict chances of accident occurrences for the Republic of Korea Air Force (ROKAF) proactively. Systematic and engineered analytical methods i.e. artificial neural network (ANN) and logistics regression are employed in practice to develop prediction models in order to predict accidents for the purpose of identifying superior technique among the two. After experimentation, it is revealed that ANN outperforms logistic regression technique in terms of enhanced prediction rate.