A carregar...
Publicação
Time Series Intervention Analysis of Portuguese Airports' Passenger Throughput (2004-2018)
| Nome: | Descrição: | Tamanho: | Formato: | |
|---|---|---|---|---|
| 8.11 MB | Adobe PDF |
Autores
Orientador(es)
Resumo(s)
The evolution of passenger traffic is shaped by economics, demographics, operating
environment and unexpected events. Understanding the dynamics of these anomalous events
is essential for accurate forecasting and informed decision-making in civil aviation. This
dissertation focuses on modelling a monthly time series with interventions, representing the
number of passengers who embarked, disembarked, or were transferred at the main
Portuguese airports between 2004 and 2018. The primary objective is to identify model
specifications that deliver accurate forecasts, utilizing a recursive out-of-sample forecasting
approach to assess performance against data from 2019. The findings demonstrate that
incorporating interventions significantly improves the models’ fit, both in-sample and out-ofsample, when compared to standard ARIMA models. However, the inclusion of additional
regressors, such as outliers, interventions and breakpoints, did not consistently lead to better
forecasting accuracy. Furthermore, the analysis reveals that monthly data aggregation is not
ideal for detecting anomalies or assessing the impact of specific events in mature and
competitive markets. In such contexts, the rapid responses of market participants often
compensate for supply disruptions caused by the anomalous events, thereby diminishing their
detectability in aggregated data.
Descrição
Dissertation presented as the partial requirement for obtaining a Master's degree in Statistics and Information Management, specialization in Information Analysis and Management
Palavras-chave
ARIMA Regression with ARIMA errors Time series outliers Intervention analysis SDG 8 - Decent work and economic growth SDG 9 - Industry, innovation and infrastructure SDG 11 - Sustainable cities and communities SDG 13 - Climate action