PREDICTIVE MODELS FOR AIR SHOW TICKET SALES

Abstract

Promoters and companies often want to forecast the success of their event while their tickets are still on sale. Prediction and good estimation of ticket sales will allow companies to see if they need to plan for things in advance, such as advertisements for their event and the arrangements of services for the event. The purpose of this study is to predict the ticket sales for air shows put on by a ticket-selling company. This study uses the ticket sales data of four past events of the Great Tennessee Air Show. We study two kinds of prediction models, the three-segment latent class Weibull model and dynamic neural network model. We then compare the prediction results of both models. We conclude that the three-segment latent classWeibull model is more suitable for long-term prediction, and the dynamic neural network model is better for short-term forecasting.