In air travel, an itinerary is a direct flight or sequence of connecting flights between two cities. The objective of itinerary market share estimation is to forecast market shares of competing itineraries. This paper examines and compares three different methods for itinerary market share estimation: multinomial logit models, artificial neural networks, and a custom model developed by the authors. Using real-world booking data, each model is constructed and calibrated to best reproduce the given data. The resulting models are applied to test data and the custom model was found to show the best results. Although multinomial logit model are used by many airlines for planning and forecasting purposes, such methods resulted in the lowest forecasting quality.
Additional information:
Das Dokument wird vom Publikationsserver der Universitätsbibliothek Mannheim bereitgestellt.
![[img]](https://madoc.bib.uni-mannheim.de/style/images/fileicons/application_pdf.png)
Search Authors in