SIGMA JOURNAL OF ENGINEERING AND NATURAL SCIENCES, cilt.42, sa.2, ss.555-565, 2024 (ESCI)
Modeling the number of air passengers correctly is essential for management policy in the
global world. Based on seasonality (depending on the season of the year), data about the
number of air passengers are heteroscedastic. Heteroscedasticity violates “Homoscedasticity” which is one of the central assumptions of linear regression analysis. In this study, a new
weighting approach called “Weighting Absolute Centered External Variable” (WCEV) is applied to the Turkish total monthly air passenger’s data to obtain correct statistical inference
and forecasting. Besides scatter plot months vs. studentized residuals, the homoscedasticity assumption is checked with the studentized RCEV test as well. Consequently, the WCEV
method is shown superior performance against multiple linear regressions and exponential
weighted moving average (EWMA) methods. The study also provides insights into the seasonal patterns of air passenger demand in Turkey, with passenger mobility increasing in the
last quarter of each year and the lowest demand in January and February. This information
can be used to optimize airport and airplane maintenance schedules and increase capacity
during peak months.