Prediction of Aeroplane Crash Fatalities Using Regularization Regression


  • Khuneswari A/P Gopal Pillay
  • Audrey Err Kiat Joo


LASSO Regression, Logistic Regression, Prediction Of Aeroplane Crash Fatalities


Air transportation is extensively used these days, and the safety of air transportation is affected as the number of aeroplane crash is getting an increase, it is important to decrease the risk of an aeroplane crash. Hence, this study aims to describe aeroplane crash fatalities based on the factors affecting aeroplane crash fatalities, compare the model selection between logistic regression and LASSO logistic regression in terms of prediction of the presence of aeroplane crash fatalities, and identify the main factors affecting aeroplane crash fatalities.  The number of aeroplane crash fatalities from 1st January 2000 to 31st December 2019 is described by using a bar chart. The performance of model selection by using LASSO logistic regression and logistic regression is compared by using the accuracy and precision obtained from the confusion matrix and the AUC value obtained from the ROC curve. The factors that affect the presence of aeroplane crash fatalities are determined from the best model. Based on the results, LASSO logistic regression showed a better performance compared to logistic regression in the analysis of prediction of aeroplane crash fatalities. In conclusion, three main factors which are the flight phase where aeroplane crash happened, regions of aeroplane crash happened, and causes of an aeroplane crash were concluded to show a significant sign that affects the presence of aeroplane crash fatalities. With this, the airlines should take more precautions to prevent the presence of aeroplane crash fatalities.




How to Cite

A/P Gopal Pillay, K., & Kiat Joo, A. E. (2021). Prediction of Aeroplane Crash Fatalities Using Regularization Regression. Enhanced Knowledge in Sciences and Technology, 1(2), 98–108. Retrieved from