Page 1 of 12
Archives of Business Research – Vol. 12, No. 12
Publication Date: December 25, 2024
DOI:10.14738/abr.1212.17968.
Phuong, N. T. T., Nhan, V. T., & Phan, V. T. (2024). An Improved the Prediction Accuracy of the Nonlinear Grey Bernoulli Model by
Fourier Series and Its Application in Container Throughput Forecasting in Danang Port. Archives of Business Research, 12(12). 01-
12.
Services for Science and Education – United Kingdom
An Improved the Prediction Accuracy of the Nonlinear Grey
Bernoulli Model by Fourier Series and Its Application in
Container Throughput Forecasting in Danang Port
Nguyen Thi Thuy Phuong
Central Region College of Technology,
Economics and Water Resources
Vu Thanh Nhan
Vietnam-Korea University of Information and
Communications Technology – University of Danang
Van Thanh Phan
Vietnam-Korea University of Information and
Communications Technology – University of Danang
ABSTRACT
In order to improve the prediction accuracy of Nonlinear Grey Bernoulli Model
NGBM (1,1), this study using Fourier series to modify their residual error of this
model. To verify the effectiveness of the proposed approach, the annual water
consumption in Wuhan from 2005 to 2012 is used for the modeling to forecast the
annual water consumption demand from 2013 to May 2020, and the forecasting
results proved that the Fourier- NGBM (1,1) is a better than the among forecasting
model used in this situation. Furthermore, this proposed approach is applied the
real case in forecasting the Container Throughput Forecasting in Danang Port. The
empirical results show that the proposed model will get a higher accuracy
performance with the lowest MAPE =1.93%. This result is not only show the
effectiveness of proposed model but also offers valuable insights for Danang
policymakers in orientation and planning management agency so as to boost the
development of upcoming port activities.
Keywords: Nonlinear Grey Bernoulli model, Fourier series, Forecasting, Accuracy,
Container, Danang Port.
INTRODUCTION
Grey forecasting is one of main part of Grey system theory, an effective method for modeling
and forecasting small sample time series. In the early 1980s, Deng [1, 2] proposed the grey
model GM (1,1) based on control theory, which is the core model used in the grey forecasting
model. This model utilizes an operator obtained by first -order accumulation to operate on the
non-negative original sequence. It demonstrates the approximate exponential growth laws and
achieves short-term forecasting accuracy. With Its advantages in dealing with uncertain
information and using as few as four data points [3,5], The GM (1,1) has been validated and
widely used in various fields such as tourism [6, 7], transportation [8- 10], financial and
economic [11- 13], integrated circuit industry [14-17], energy industry [18-20] etc...
Page 2 of 12
2
Archives of Business Research (ABR) Vol. 12, Issue 12, December-2024
Services for Science and Education – United Kingdom
In the recent years, there are many scholars propose new procedures with different ways to
improve the precision accuracy of GM (1,1) model. For instant, Lin et al. [21] and Wang et al.
[22] used different methods to calculate new background values to improve the background
values. Hsu [17] and Wang et al. [23] used different methods to modified internal parameter
estimation like development coefficient and grey input coefficient. Some scholars have
established GM (1,1) model with residuals modification like Hsu [15] and Wang et al. [24]. In
addition, many hybrid models based on GM (1,1) were proposed. These include the grey
econometric model [25], the grey Markov model [27, 28], and the grey fuzzy model [21], etc.
Despite its improvement in prediction accuracy, the prediction accuracy of the GM (1,1) model
is always monotonic. As a result, GM (1,1) model may not be always satisfactory. The recently
developed nonlinear grey Bernoulli model NGBM (1,1) is a new grey forecasting model [28]. It
has a power exponent n that can effectively manifest the nonlinear characteristics of real
systems and flexibility determines the shape of the model’ curve. Unlike GM (1,1) and the grey
Verhulst model which rely on a constant number such as 0 or 2, the NGBM (1,1) does not
require such a number (excluding 1).
Therefore, forecasting of the fluctuation sequence can be performed by the fluctuation features
as long as the power exponent and structural parameters in the model are known. The NGBM
(1,1) was successfully used to simulate and forecast the values of the annual unemployment
rates of ten selected countries and foreign exchange rates of Taiwan’s major trading partners
[29, 30]. This success indicates that the NGBM (1,1) significantly improves the accuracy of the
simulation and forecasting predictions of the original GM (1,1). Zhou et al. [30] selected the
value of n by using a particle swarm optimization algorithm and used the model to forecast the
power load of the Hubei electric power network. Hsu [16] used the genetic algorithm to
optimize parameters of the NGBM (1,1) and applied it to forecast the economic trends in the
integrated circuit industries in Taiwan. Chen et al. [31] proposed a Nash NGBM (1,1) in which
an interpolated coefficient in the background value is introduced into the NGBM (1,1) and the
parameters solved for based on the Nash equilibrium concept. This strengthens the adaptability
of the model towards the original data and eventually improves the accuracy of the model.
Later, Wang, et al. [32] proposed optimized NGBM (1,1) model to forecast the qualified
discharge rate of the industrial waste water in 31 administrative areas in China by improved
background interpolation value p and exponential value n is put forward in an NGBM (1,1). Pao
et al. [33] forecasted CO2 emissions, energy consumption, and economic growth in China based
on ARIMA and NGBM (1,1). Performance evaluation results showed that the NGBM (1,1) can be
used safely for future projection of these indicators in clean energy economy.
All these improvements focus on the model parameters and the background value. In fact, the
initial condition is also an important factor determining the grey modeling accuracy. This is
because the initial condition is a part of the predictive function. The current paper aims to
develop an approach to increase the predictive precision of the NGBM (1,1) by modifying the
residual error obtained from NGBM (1,1) with Fourier series. Numerical example and practical
application shows that the proposed Fourier-NGBM (1,1) model has higher performance on
models prediction. The remainder of this paper is organized as follows. A brief introduction to
the original NGBM (1,1) and the method of optimization of the modified residual error by
Fourier series are given in Section 2. Section 3 proves the effectiveness of using the proposed
Fourier-NGBM (1,1) by comparison with the Coupling model of Grey system and Multivariate