Predicting the Number of Fatalities in an Air Crash

Air travel is one of the safest forms of travel today. The mortality rate in air travel is much lesser than that in automobile accidents. Nevertheless, aviation accidents have always been a huge cause of concern for people. Crashes could occur due to multiple factors like injury severity, weather condition, amateur built, make, etc. Air crashes could result in fatalities or serious injuries. The outcome of this research is to predict the number of fatalities when factors like the ones mentioned above have been provided. Air travel has a wealth of historical data available, and by analyzing patterns in flight accidents, there is potential to enhance the safety of air travel further, when compared to other modes of transport. This is essential as this can help the rescue team to respond accordingly. Three models namely, Gradient-boosting-regression, Random-Forest-Regressor, XGBoost have been implemented. After reviewing the results of all three models it has been found that Random-Forest comparatively better results.