Minimum covariance determinant-based bootstrapping for appraising air passenger arrival data

Abstract

Air travel management is a case-special process since it includes different types of uncertainties such as ungovernable passenger mobility, variable costs as well as extraordinary restrictions like the Covid-19 pandemic. Therefore, the use of robust and reproducible statistical evaluations under uncertainty is required. The cornerstone of this study is the adaptation of bootstrapping and the robust Minimum Covariance Determinant (MCD)-based parameter estimation under a heterogeneous process. In addition, the study includes a novel bootstrapping regression implementation. The methodological developments have been tested by Serbia's air transport data. The results showed that combining robust estimator and bootstrapping provides some advantages for determining outliers and also making advanced diagnostics. Thus, a state-of-the-art approach based on accuracy, reproducibility, and transparency has been introduced and its usability in the air travel mobility process has been exhibited.