Notice of Retraction: Airline demand forecast based on panel data model

Abstract

Airline demand forecast is a very important task for air companies to operate an existing airline or open a new airline. In this paper we introduce panel data model to forecast airline demand that gives consideration to both advantages of time series method and cross sectional regression method, which takes the specific characteristic of each individual airline into account. We construct four demand forecasting models by classifying Flying Range and Ticket Price and get function expression for each model. We find that when Flying Range is less than 1000 km and the Ticket Price is lower than 1000 Y, the Airline Demand is mainly subject to Ground Traffic and the Airline Demand of current period is probably affected by the one of prior period, otherwise, both independent variables of Gross Region Product and Ground Traffic have significant positive effects to Airline Demand. Lastly we use the constructed models to forecast some airline demands in 2006 and the results show that the models are well predictable and satisfactory.