Developing intuitionistic fuzzy seasonality regression with particle swarm optimization for air pollution forecasting
| Published date | 08 April 2019 |
| Date | 08 April 2019 |
| Pages | 561-577 |
| DOI | https://doi.org/10.1108/IMDS-02-2018-0063 |
| Author | Chung-Han Ho,Ping-Teng Chang,Kuo-Chen Hung,Kuo-Ping Lin |
| Subject Matter | Information & knowledge management,Information systems,Data management systems,Knowledge management,Knowledge sharing,Management science & operations,Supply chain management,Supply chain information systems,Logistics,Quality management/systems |
Developing intuitionistic fuzzy
seasonality regression with
particle swarm optimization
for air pollution forecasting
Chung-Han Ho
Department of Industrial Management,
National Taiwan University of Science and Technology, Taipei, Taiwan
Ping-Teng Chang
Department of Industrial Engineering and Enterprise Information,
Tunghai University, Taichung, Taiwan
Kuo-Chen Hung
Department of Computer Science and Information Management,
Hungkuang University, Taichung, Taiwan, and
Kuo-Ping Lin
Institute of Innovation and Circular Economy, Asia University, Taichung, Taiwan
Abstract
Purpose –The purpose of this paper is to develop a novel intuitionistic fuzzy seasonality regression (IFSR)
with particle swarm optimization (PSO) algorithms to accurately forecast air pollutions, which are typical
seasonal time series data. Seasonal time series prediction is a critical topic, and some time series data contain
uncertain or unpredictable factors. To handle such seasonal factors and uncertain forecasting seasonal time
series data, the proposed IFSR with the PSO method effectively extends the intuitionistic fuzzy linear
regression (IFLR).
Design/methodology/approach –The prediction model sets up IFLR with spreads unrestricted so as to
correctly approach the trend of seasonal time series data when the decomposition method is used. PSO
algorithms were simultaneously employed to select the parameters of the IFSR model. In this study, IFSR
with the PSO method was first compared with fuzzy seasonality regression, providing evidence that the
concept of the intuitionistic fuzzy set can improve performance in forecasting the daily concentration of
carbon monoxide (CO). Furthermore, the risk management system also implemented is based on the
forecasting results for decision-maker.
Findings –Seasonal autoregressive integrated moving average and deep belief network were then employed
as comparative models for forecasting the daily concentration of CO. The empirical results of the proposed
IFSR with PSO model revealed improved performance regarding forecasting accuracy, compared with the
other methods.
Originality/value –This study presents IFSR with PSO to accurately forecast air pollutions. The proposed
IFSR with PSO model can efficiently provide credible values of prediction for seasonal time series data in
uncertain environments.
Keywords Intuitionistic fuzzy linear technology, Intuitionistic fuzzy seasonality regression,
Seasonal time series data
Paper type Research paper
1. Introduction
Seasonal time series of a regular nature are encountered in many fields, such as soil dryness
index (Li et al., 2003), tourism demand (Huang and Min, 2002), municipal solid waste
management (Navarro-Esbrí et al., 2002) and electricity prices (Cerqueti et al., 2017).
The seasonal autoregressive integrated moving average (SARIMA) model is the best-known
approach and was introduced by Box and Jenkins (1976). In recent years, the artificial neural
Industrial Management & Data
Systems
Vol. 119 No. 3, 2019
pp. 561-577
© Emerald PublishingLimited
0263-5577
DOI 10.1108/IMDS-02-2018-0063
Received 6 February 2018
Revised 19 July 2018
Accepted 22 September 2018
The current issue and full text archive of this journal is available on Emerald Insight at:
www.emeraldinsight.com/0263-5577.htm
561
Developing
IFSR
network (ANN) model has also been applied to forecast seasonal data patterns (Zhang and
Qi, 2005). The literature (Hamzaçebi, 2008; Nam and Schaefer, 1995; Pai et al., 2010;
Williams, 1997) indicates that ANN can obtain desirable results in seasonal and trend
forecasting. However, although this model has the advantage of accurate forecasting in
short periods, it has several limitations. First, clear and definite functional relationships with
current values, past values and white noise should be assumed. Second, the numbers of
observations that are required should be at least 50 and preferably 100 or more (Tseng et al.,
1999). Finally, uncertain or unpredictable factors should be contained.
Fuzzy regression(FR) (Tanaka et al., 1980) is a methodologythat can handle unpredictable
factors or uncertainties in time series data. Kim et al. (1996) noted that classical regression
analysis uses rigid assumptions about statistical properties. FR may relax these rigid
assumptions regarding properties such as the normality of error terms and predictions, as
well as random measurement errors in recorded observations. When this approach is used,
observationaluncertainties or fuzzinessare represented by the fuzzy parametersrelated to the
indefinite structure of the system.The objective function then minimizes the total fuzziness of
the estimated outputs yielded based on the condition that an H-level-set inclusion of all
observed data holds (Tanaka et al., 1980). The various FR methods can be seen in the
following:Bisserier et al. (2010), Chan et al. (2010),Chang and Lee (1994a,b), Chang et al. (2009),
D’Urso et al. (2010, 2011), Kacprzyk and Fdrizzi (1992), Lee et al. (2010), Lin and Pai (2010),
Sakawa and Yano (1992), Tanaka et al. (1989), Watada et al. (2009), Fan et al. (2013), Nureize
et al. (2014),Jiang et al. (2017). Chang and Lee (1994a,b) proposed a concept of FR analysisthat
allows for the possibility that spreadsof the fuzzy parameterscan be unrestricted parameters.
This overcomes the problem that both the trends of modes and the spreads of FR may be
inconsistent and conflicting and, if not treated properly, may predict incorrect trends.
The unrestrictedFR may correct the trends of the predictedmodel. Chang (1997) appliedFR to
a fuzzy forecasting technique for the seasonality of time series data. The fuzzy forecasting
technique can analyze both seasonal fuzziness and model trends. Parvathi et al. (2013)
proposed an intuitionistic fuzzy linear regression (IFLR), with which they attempted to
minimize the total fuzziness of the model, which is related to the width of the intuitionistic
fuzzy set (IFS) coefficient. These previous studies successfully extended the fuzzy linear
regression of Tanaka et al. (1980). Nureize et al. (2014) developed a fuzzy random multi-
attribute evaluation model with confidence intervals using expectations and variances of
fuzzy random variables. They revealed flaws in traditional FR and, as a result, successfully
modified or extended traditional FR. Based on the views expressed in previous research,
which includes studies on unrestricted FR, fuzzy seasonality forecasting, and intuitionistic
FR models, the IFLRmethod should be extended to include the seasonality of time seriesdata.
Therefore,this study develops an intuitionisticfuzzy seasonality regression(IFSR) model that
combines the decomposition techniques and intuitionistic FR to forecast seasonal time series
patterns. In addition, to obtain the correct trend, intuitionistic FR with unrestricted spreads
has also been developed.Furthermore, the parametersof traditional FR are typicallyprovided
by expert knowledge, which may be less easy to obtain in different cases. In this study,
particle swarm optimization (PSO) was employed to identify the optimal performance of the
proposed IFSR, which can obtain optimal parameters of the proposed IFSR model.
Air pollution is a global environmental and health issue that impacts people’s thoughts and
experiences in their lives directly through visual perceptions. Air pollution characteristics, which
can change dramatically based on time, space, weather and climate, are not normally collected in
surveys, but instead from monitor stations (Luechinger, 2009; Schmitt, 2013). Li et al. (2018)
proposed using psychophysics application to quantify air pollution’simpactonsubjective
well-being, which verified that air pollution does influence people’s thoughts. The challenge of
managing air pollution is significant, because of the many types of air pollutants, insufficient
funds for monitoring and abatement programs, and political and social challenges in defining
562
IMDS
119,3
To continue reading
Request your trial