Tourism Time Series Forecast

Tourism Time Series Forecast

João Paulo Teixeira (Polytechnic Institute of Bragança, Portugal) and Paula Odete Fernandes (Polytechnic Institute of Bragança, Portugal)
DOI: 10.4018/978-1-4666-8368-6.ch005
OnDemand PDF Download:
No Current Special Offers


In this chapter four combinations of input features and the feedforward, cascade forward and recurrent architectures are compared for the task of forecast tourism time series. The input features of the ANNs consist in the combination of the previous 12 months, the index time modeled by two nodes used to the year and month and one input with the daily hours of sunshine (insolation duration). The index time features associated to the previous twelve values of the time series proved its relevance in this forecast task. The insolation variable can improved results with some architectures, namely the cascade forward architecture. Finally, the experimented ANN models/architectures produced a mean absolute percentage error between 4 and 6%, proving the ability of the ANN models based to forecast this time series. Besides, the feedforward architecture behaved better considering validation and test sets, with 4.2% percentage error in test set.
Chapter Preview

1. Introduction

Several studies have been published in the field of tourism in recent years. The forecast of tourism demand has an important role to play in the planning, decision-making and controlling of the tourism sector (Witt & Witt, 1995; Wong, 2002; Fernandes, 2005; Yu & Schwartz, 2006).

Tourism has been seen as an important sector in the Portuguese economy. Decision makers should adopt policies that ensure its profitability and sustainability, (Dolgner & Costa, 2010). It has a strategic importance for the Portuguese economy because of its ability to create wealth and employment. This sector shows competitive advantages as few other, (Ministério da Economia e da Inovação, 2006).

According to the World Tourism Organization (WTO), Portugal will achieve 18.3 million foreign visitors in 2020. The population of Portugal is about 10 million. Tourism is at present one of Portugal’s most important activities. Apart from its impact on the balance of payments and Gross Domestic Product (GDP), and its role on employment generation, investment and revenue, it is also recognized as an “engine” for development and other economic activities, (World Tourism Organization, 2011).

The incomes of the foreign tourism can be seen as exportations which increases the importance of this activity because it contributes to the equilibrium of the import/export balance. Additionally, if the Portuguese tourism marketing convinces the Portuguese people to spend their vacancies inside the country it will reduce the exportations, therefore reinforcing again the import/export balance equilibrium.

The Northern region of Portugal is a very particular region that offers an interesting alternative to the so called 'mass tourism', focusing on the provision of a wide variety of tourism products that range from the beach, mountains, thermal/health spas, gastronomy and rural tourism, which has had a significant increase in recent years, (Fernandes, 2005).

In this respect, and given the substantial growth of this sector in the North of Portugal, the development of models is needed to make reliable forecasts of tourism demand. These forecasting models assume an important role for the process of planning and decision-making both within the public and private sectors.

Currently there are a wide range of methods that have emerged in response to most situations, displaying different characteristics and methodologies ranging from the simplest linear regression model to more complex approaches. The Box-Jenkins forecasting models belong to the family of algebraic models known as ARIMA models, which make it possible to make forecasts based on a given stationary time series. The methodology considers that a real time series amounts to a probable realization of a stochastic process. The aim of this analysis is to identify the model that best depicts the underlying unknown stochastic process and a good representation of its realization, i.e. of the real-time series. The ANN methodology also has had countless applications in the most diverse areas of knowledge and has been used in the field of forecasting as an alternative to classical models.

Complete Chapter List

Search this Book: