TAIEX forecasting based on fuzzy time series, particle swarm optimization techniques and support vector machines
Introduction
Fuzzy time series has successfully been used to deal with various forecasting problems, such as to forecast the temperature, the economy, the inventory, the earthquake, and the stock index. In [28], Song and Chissom proposed the concepts of fuzzy time series. In [29], [30], Song and Chissom presented the time-invariant fuzzy time series model and the time-variant fuzzy time series model to forecast the enrollments of the University of Alabama. In recent years, some fuzzy forecasting methods based on fuzzy time series have been presented [2], [4], [5], [6], [7], [8], [9], [10], [11], [12], [13], [14], [15], [16], [17], [18], [19], [20], [21], [23], [24], [25], [26], [27], [31], [32], [33], [34], [35], [36]. In [2], Chen and Chen presented a method used multivariate to forecast the Taiwan Stock Exchange Capitalization Weighted Stock Index (TAIEX) [39]. In [4], Chen presented a high-order fuzzy time series model for forecasting the enrollments of the University of Alabama to get higher forecasting accuracy rates. In [6], Chen and Chen presented a method to forecast the TAIEX by fuzzy variation groups. In [7], Chen and Chung presented a method for forecasting the enrollments of the University of Alabama using fuzzy time series and genetic algorithms. In [8], Chen and Hsu presented a method to forecast the enrollments of the University of Alabama based on fuzzy time series. In [9], Chen and Hwang presented a two-factors time-variant fuzzy time series model to predict the temperature. In [11], Chen et al. presented a method for fuzzy forecasting based on two-factors high-order fuzzy-trend logical relationship groups and particle swarm optimization techniques. In [12], Chen and Wang presented a high-order fuzzy time series forecasting method using fuzzy-trend logical relationships. In [13], Chen and Wang presented a method for forecasting the enrollments of the University of Alabama by using automatic clustering techniques and fuzzy logical relationships. In [15], Huarng observed that the length of intervals in the universe of discourse can affect the forecasting results and proposed the distribution-based length method and the average-based length method for handling the forecasting problems. In [16], Huarng presented a method using a heuristic function to forecast the enrollments of the University of Alabama and the Taiwan Futures Exchange (TAIFEX). In [17], Huarng and Yu presented a method using type-2 fuzzy time series to forecast the TAIEX. In [36], Teoh et al. presented a method to forecast the stock markets based on a hybrid multi-order fuzzy time series model, which employed the rough set theory to mine fuzzy logical relationships from time series. In [14], Hung and Lin developed a novel intuitionistic fuzzy least-squares support vector regression with genetic algorithms to forecast the long-term indexes of business cycles. In [26], Lin et al. developed a fuzzy least-squares support vector regression model with genetic algorithms to forecast seasonal revenues. However, because the forecasting accuracy rates of the existing methods are not good enough, a new method is needed to be developed to overcome the drawback of the existing methods to get higher forecasting accuracy rates.
In this paper, a new method for forecasting the TAIEX is presented based on fuzzy time series, particle warm optimization techniques [22] and support vector machines [37]. The proposed method to forecast the TAIEX is based on the slope of one-day variation of the TAIEX and the slope of two-days average variation of the TAIEX. Because the slope of two-days average variation of the TAIEX is smoother than the slope of one-day variation of the TAIEX, it is chosen to define the universe of discourse. The particle swarm optimization techniques are used to get optimal intervals in the universe of discourse. The support vector machine is used to classify the training data set. The first feature and the second feature of the support vector machine are the slope of one-day variation and the slope of two-days average variation of the TAIEX, respectively. The experimental results show that the proposed method outperforms the existing methods for forecasting the TAIEX.
Section snippets
Fuzzy time series
In this section, some basic concepts of fuzzy time series are reviewed from [3], [28], [29], [30], where the values of fuzzy time series are represented by fuzzy sets [38]. Let U be the universe of discourse, where U = {u1, u2, … , un}. A fuzzy set Ai in the universe of discourse U can be represented bywhere is the membership function of the fuzzy set Ai, denotes the degree of membership of uj belonging to the fuzzy set Ai, [0, 1] and 1 ⩽ j ⩽ n.
Particle swarm optimization
In [22], Kennedy and Eberhart developed an optimization algorithm, named particle swarm optimization (PSO), which was inspired by the social behavior of bird flocking or fish schooling. In PSO, a set of particles consists of a particle swarm, where a particle denotes a potential solution. The position and the velocity of the ith particle in an n-dimensional search space can be represented by Xi = [xi,1, xi,2, … , xi,n] and Vi = [vi,1, vi,2, … , vi,n], respectively. Each particle has the personal best
A new method for forecasting the TAIEX based on fuzzy time series, particle swarm optimization techniques and support vector machines
In this section, a new method to forecast the TAIEX is presented based on fuzzy time series, particle swarm optimization techniques and support vector machines. The proposed method is now presented as follows:
Step 1: Let V1t denote the variation between trading day t and trading day t − 1 and let V2t denote the average variation between trading day t and trading day t − 2, respectively, shown as follows:where TAIEXt−2, TAIEXt−1 and TAIEXt denote the TAIEX
Experimental results
In this section, the proposed method is applied to forecast the TAIEX from 1990 to 2004. In order to compare the experimental results of the proposed method with the ones of the methods presented in [3], [5], [6], [10], [16], [18], [24], [30], [31], a 10 month/2 month split for training/testing is also adopted, i.e., for each year, the data from January to October are used as the training data set, and the data from November to December are used as the testing data set. The Root Mean Squared
Conclusions
In this paper, a new method for forecasting the TAIEX is presented based on fuzzy time series, particle swarm optimization techniques and support vector machines. The proposed method forecast the TAIEX based on slope of one-day variation of the TAIEX and the slope of two-days average variation of the TAIEX. Because the slope of two-days average variation of the TAIEX is smoother than the slope of one-day variation of the TAIEX, it is chosen to define the universe of discourse. The particle
Acknowledgement
This work was supported in part by the National Science Council, Republic of China, under Grant NSC 101-2221-E-011-171-MY2.
References (39)
Forecasting enrollments based on fuzzy time series
Fuzzy Sets and Systems
(1996)- et al.
Multi-variable fuzzy forecasting based on fuzzy clustering and fuzzy interpolation techniques
Information Sciences
(2010) - et al.
Forecasting enrollments using automatic clustering techniques and fuzzy logical relationships
Expert System with Applications
(2009) - et al.
Long-term business cycle forecasting through a potential intuitionistic fuzzy least-squares support vector regression approach
Information Sciences
(2013) Effective lengths of intervals to improve forecasting in fuzzy time series
Fuzzy Sets and Systems
(2001)Heuristic models of fuzzy time series for forecasting
Fuzzy Sets and Systems
(2001)- et al.
A type 2 fuzzy time series model for stock index forecasting
Physica A
(2005) - et al.
The application of neural networks to forecast fuzzy time series
Physica A
(2006) - et al.
Forecasting TAIFEX based on fuzzy time series and particle swarm optimization
Expert Systems with Applications
(2010) - et al.
An improved method for forecasting enrollments based on fuzzy time series and particle swarm optimization
Expert Systems with Applications
(2009)
Revenue forecasting using a least-squares support vector regression model in a fuzzy environment
Information Sciences
A comparison of fuzzy forecasting and Markov modeling
Fuzzy Sets and Systems
Fuzzy time series and its model
Fuzzy Sets and Systems
Forecasting enrollments with fuzzy time series—Part I
Fuzzy Sets and Systems
Forecasting enrollments with fuzzy time series—Part II
Fuzzy Sets and Systems
Temperature prediction and TAIFEX forecasting based on automatic clustering techniques and two-factors high order fuzzy time series
Expert Systems with Applications
A bivariate fuzzy time series model to forecast the TAIEX
Expert Systems with Applications
Corrigendum to “A bivariate fuzzy time series model to forecast the TAIEX
Expert Systems with Applications
A hybrid multi-order fuzzy time series for forecasting stock markets
Expert Systems with Applications
Cited by (165)
Aggregation and definition of an algebraic framework for fuzzy time series: An application in the supply-demand domain
2022, International Journal of Approximate ReasoningA fuzzy regression functions approach based on Gustafson-Kessel clustering algorithm
2022, Information SciencesParticle swarm optimization of partitions and fuzzy order for fuzzy time series forecasting of COVID-19
2021, Applied Soft ComputingCitation Excerpt :Kuo et al. [22] have proposed a new hybrid model to provide forecasting solution based on first order and higher order FTS combined with PSO. Chen and Kao [23] have proposed a new FTS forecasting method for Taiwan Stock Exchange Capitalization Weighted Stock Index (TAIEX) forecasting. The authors have defined UOD for TAIEX using average variation of slope of two-days because it is smoother compared to the single-day variation slope.
Picture fuzzy set and quality function deployment approach based novel framework for multi-criteria group decision making method
2021, Engineering Applications of Artificial IntelligenceImproving fuzzy clustering algorithm for overlapping elements and its application
2024, International Journal of Information Technology (Singapore)