Genetic algorithm based fuzzy time series tourism demand forecast model
| Published date | 11 April 2016 |
| Pages | 483-507 |
| Date | 11 April 2016 |
| DOI | https://doi.org/10.1108/IMDS-05-2015-0165 |
| Author | Sumit Sakhuja,Vipul Jain,Sameer Kumar,Charu Chandra,Sarit K Ghildayal |
| Subject Matter | Information & knowledge management,Information systems,Data management systems |
Genetic algorithm based fuzzy
time series tourism demand
forecast model
Sumit Sakhuja
Mechanical Engineering Department,
Indian Institute of Technology Delhi, New Delhi, India
Vipul Jain
Department of Industrial Engineering and Management,
University of Sharjah, Sharjah, United Arab Emirates
Sameer Kumar
Department of Operations and Supply Chain Management,
University of St Thomas, Opus College of Business,
Minneapolis, Minnesota, USA
Charu Chandra
Dearborn College of Business Administration,
University of Michigan –Dearborn, Dearborn, Michigan, USA, and
Sarit K. Ghildayal
Department of Computer Science, University of Minnesota,
Minneapolis, Minnesota, USA
Abstract
Purpose –Many studies have proposed variant fuzzy time series models for uncertain and vague
data. The purpose of this paper is to adapt a fuzzy time series combined with genetic algorithm (GA) to
forecast tourist arrivals in Taiwan.
Design/methodology/approach –Different cases are studied to understand the effect of variation of
fuzzy time series order, number of intervals and population size on the fitness function which decreases
with increase in fuzzy time series order and number of fuzzy intervals, but do not have marginal effect
due to change in population size.
Findings –Results based on an example of forecasting Taiwan’s tourism demand was used to verify
the efficacy of proposed model and confirmed its superiority to existing models providing solutions for
different orders of fuzzy time series, number of intervals and population size with a smaller forecasting
error as measured by root mean square error.
Originality/value –This study provides a viable forecasting methodology, adapting a fuzzy time
series combined with an evolutionary GA. The proposed hybridized framework of fuzzy time series
and GA, where GA is used to calibrate fuzzy interval length, is flexible and replicable to many
industrial situations.
Keywords Decision support system, Evolutionary algorithm, GA,
Adaptive fuzzy time series forecasting model, Interval calibration
Paper type Research paper
1. Introduction
In the past few decades, tourism has clearly become one of the most prominent
economic trends for many countries. The tourism industry requires accurate forecasts
in order to build and manage its service supply chains efficiently. The need for more
Industrial Management & Data
Systems
Vol. 116 No. 3, 2016
pp. 483-507
©Emerald Group Publis hing Limited
0263-5577
DOI 10.1108/IMDS-05-2015-0165
Received1May2015
Revised 21 July 2015
20 September 2015
Accepted 24 September 2015
The current issue and full text archive of this journal is available on Emerald Insight at:
www.emeraldinsight.com/0263-5577.htm
483
Tourism
demand
forecast model
accurate forecasts of tourism demand is driven by the desire to reduce risk and
uncertainty. This need, according to Frechtling (2001), is further stressed by the
perishable nature of tourism products, the simultaneous purchase-production process,
and the role of complementary services in shaping consumer satisfaction and, thereby,
its subsequent demand for future tourism services. Government bodies need accurate
tourism demand foreca sts to plan required touri sm infrastructures , such as
accommodation, site planning and transportation development. Tourism forecasting
is classified into three main groups according to the methods and techniques
adopted –an econometric-based approach, time series techniques and artificial
intelligence (AI)-based methods. Before the 1990s, traditional regression approaches
dominated the tourism forecasting and modeling literature. After pioneering up-to-date
developments in econometric methodologies in recent years, the reputation of
econometric forecasting models for improved accuracy has grown (Song and Li, 2008).
Time series models such as ARIMA and GARCH (Alleyne, 2006; Gil-Alana et al., 2004;
Lee et al., 2008; Lim and McAleer, 2002) and econometric models such as error
correction model (ECM) and the vector autoregressive (VAR) models (Song and
Witt, 2006; Wong et al., 2007) are the most commonly used tourism demand forecasting
techniques. The single exponential smoothing (SES) model is used to forecast a time
series when there is no trend or seasonal pattern. According to Chen et al. (2008), SES is
more suitable for a time series with seasonality removed. Other studies advocate that
the ARIMA and SARIMA (Seasonal ARIMA) approaches are favored in tourism
demand forecasting when the time series does not demonstrate structural breaks
(Chu, 2008; Gustavsson and Nordström, 2001). Preez and Witt (2003) show tha t the
ARIMA approach performs best in terms of forecasting accuracy and goodness
of fit. Guo (2007) employs the gravity model to analyze inbound tourism demand to
China, and Khadaroo and Seetanah (2008) use it to examine the effect of transportation
infrastructure on tourism flows. According to Wang (2004), AI forecasting methods,
including neural networks, rough sets theory, fuzzy time series theory, grey theory,
genetic algorithms (GAs), and expert systems, tend to perform better than traditional
forecasting methods. In traditional forecasting, piecewise linear function are the
basic elements of the prediction model and users need to specify the functional form
of the problem. Obtaining proper and valid modelsaredonethrough experimentation
with possible function relations and algorithms, and such experiments take
comparatively longer time. Some popular techniques to solve the complex
engineering and optimization problems (Konar, 2005) are AI techniques such as
artificial neural networks (ANNs), fuzzy logic, and GAs. In conventional forecasting
techniques a large amount of sample data and long-term historical data is required.
Artificial models can be used to estimate the non-linear relationship, without the
limits of traditional time series and econometric models. The new elements that are
present now are the methods and techniques of soft computing and machine learning
for decision making and forecasting, in particular neural networks and GAs, and the
new science of the combination of those tools (Leigh et al., 2002). Fuzzy time series
combined with soft computing techniques can overcome these limitations and make
appropriate short-term forecasting.
Theobjectiveofthispaperistoputforwardmethodof fuzzy time series combined with
evolutionary algorithm (GA) to calibrate the length of fuzzy intervals of the fuzzy time
series. The GA will aid in searching for the best fuzzy interval sizes corresponding to which
the forecasting accuracy or mean square error is optimum. The study also aims to analyze
the effect of different fuzzy time series order, number of intervals and population size.
484
IMDS
116,3
To continue reading
Request your trial