Machine Learning for Air Transport Planning and Management
This work addresses forecasting challenges for the air transport industry, but it is incremental as it applies existing methods to a specific domain.
The paper compared multiple machine learning algorithms, including ANN and ANFIS, against traditional multiple linear regression for modeling air transport demand, finding that ANN and ANFIS achieved the lowest mean squared error.
In this work we compare the performance of several machine learning algorithms applied to the problem of modelling air transport demand. Forecasting in the air transport industry is an essential part of planning and managing because of the economic and financial aspects of the industry. The traditional approach used in airline operations as specified by the International Civil Aviation Organization is the use of a multiple linear regression (MLR) model, utilizing cost variables and economic factors. Here, the performance of models utilizing an artificial neural network (ANN), an adaptive neuro-fuzzy inference system (ANFIS), a genetic algorithm, a support vector machine, and a regression tree are compared to MLR. The ANN and ANFIS had the best performance in terms of the lowest mean squared error.