ESTIMATING A REGIONAL AIRPORT AIR PASSENGER DEMAND USING AN ARTIFICIAL NEURAL NETWORK APPROACH: THE CASE OF HUAHIN AIRPORT, THAILAND

: Artificial neural networks (ANNs) are a promising modelling approach for predicting an airport’s air passenger demand. The study proposed and empirically tested an artificial neural network model to predict the annual passenger demand for Huahin Airport, a regional and tourist focused airport located in Thailand. The ANN input variables included Thailand’s population size, Thailand’s real GDP, world jet fuel prices, Thailand total passengers carried, Thailand’s tourist numbers and Thailand’s unemployment rates. The data were trained using the Levenberg-Marquandt back-propagation algorithm. The ANN comprises eight neurons in the hidden layer and one neuron in the output layer. 80 per cent of the data was used in the training phase with the remaining data divided into validation (10 per cent) and testing (10 per cent) phases. The proposed ANN provided very accurate prediction values. The coefficient of determination R value of model was around 0.995, and the mean absolute percentage error (MAPE) of the final ANN model was 13.27%. The study found that the four key determinants of Huahin Airport annual air passenger demand were Thailand population size, the commencement of AirAsia services at Huahin Airport, Thailand’s tourist numbers, and Thailand’s real GDP.