Fuzzy linear regression : application to the estimation of air transport demand

The demand for an air transport market is very sensitive to many factors. The vagueness of the impacts of all these factors makes the task of prediction of this demand by classical methods very hazardous, especially when this estimation is used afterwards for critical decisions such as those related with the definition of supply (frequency of flights, number of seats put on the market, trip price..). Then it appears that crisp methods are not able to take fully into account all the uncertainty making up the demand while possibilistic reasoning could be a way to catch it. Following this idea, it is shown in this communication how regressions based on fuzzy logic which combine statistics and experts' attitudes can be used to improve the estimation for air transport demand. In the first section of the communication, following Tanaka's model, fuzzy linear regression is introduced. Then in the second part an extension using trapezoidal fuzzy numbers is displayed. Finally, in the last section, the application of the proposed fuzzy linear regression to the estimation of air transport demand is considered.