Novel robust approach for constructing Mamdani-type fuzzy system based on PRM and subtractive clustering algorithm

Abstract

A novel approach for constructing robust Mamdani fuzzy system was proposed, which consisted of an efficiency robust estimator (partial robust M-regression, PRM) in the parameter learning phase of the initial fuzzy system, and an improved subtractive clustering algorithm in the fuzzy-rule-selecting phase. The weights obtained in PRM, which gives protection against noise and outliers, were incorporated into the potential measure of the subtractive cluster algorithm to enhance the robustness of the fuzzy rule cluster process, and a compact Mamdani-type fuzzy system was established after the parameters in the consequent parts of rules were re-estimated by partial least squares (PLS). The main characteristics of the new approach were its simplicity and ability to construct fuzzy system fast and robustly. Simulation and experiment results show that the proposed approach can achieve satisfactory results in various kinds of data domains with noise and outliers. Compared with D-SVD and ARRBFN, the proposed approach yields much fewer rules and less RMSE values.

Keywords

Mamdani-type fuzzy system / robust system / subtractive clustering algorithm / outlier / partial robust M-regression

References

[1] HuangW-h, FangK-l, ZhangZheng. H^∞ stability analysis of Mamdani system based on fuzzy Lyapunov function [J]. Journal of Central South University: Science and Technology, 2011, 42(1): 314-319

[2] LamH K, LiH-y, DetersC, SeccoE L, WurdemannA, AlthoeferK. Control design for interval type-2 fuzzy systems under imperfect premise matching [J]. IEEE Transactions on Industrial Electronics, 2014, 61(2): 956-968

[3] CordónO. A historical review of evolutionary learning methods for Mamdani-type fuzzy rule-based systems: Designing interpretable genetic fuzzy systems [J]. International Journal of Approximate Reasoning, 2011, 52: 894-913

[4] YenJ, WangL. Simplifying fuzzy rule-based models using orthogonal transformation methods [J]. IEEE Transactions on Systems, Man, and Cybernetics, Part B: Cybernetics, 1999, 29(1): 13-24

[5] YanarT A, AkyürekZ. Fuzzy model tuning using simulated annealing [J]. Expert Systems with Applications, 2011, 38: 8159-8169

[6] PetersonM P, BedeB, StefaniniL. Fuzzy systems of Mamdani type in the LU representation [C]. IFSA World Congress and NAFIPS Annual Meeting (IFSA/NAFIPS), 2013Edmonton, A BIEEE908-913

[7] AlcaláR, Alcalá-FdezJ, GactoM J, HerreraF. Rule base reduction and genetic tuning of fuzzy systems based on the linguistic 3-tuples representation [J]. Soft Computing, 2007, 11(5): 401-419

[8] BottaA, LazzeriniB, MarcelloniF. Context adaptation of Mamdani fuzzy rule based systems [J]. International Journal of Intelligent Systems, 2008, 23(4): 397-418

[9] GactoM J, AlcaláR, HerreraF. Integration of an index to preserve the semantic interpretability in the multi-objective evolutionary rule selection and tuning of linguistic fuzzy systems [J]. IEEE Transactions on Fuzzy Systems, 2010, 18(3): 515-531

[10] MohamadiH, HabibiJ, AbadehM S, SaadiH. Data mining with a simulated annealing based fuzzy classification system [J]. Pattern Recognition, 2008, 41(5): 1824-1833

[11] KimK, KyungK M, ParkC W, KimE, ParkM. Robust TSK fuzzy modeling approach using noise clustering concept for function approximation [J]. Lecture Notes in Computer Science: Computational & Information Science, 2004, 3314: 538-543

[12] ChuangC C, SuS F, ChenS S. Robust TSK fuzzy modeling for function approximation with outliers [J]. IEEE Transactions on Fuzzy Systems, 2011, 9(8): 810-821

[13] LeskiJ M. Epsilon-insensitive fuzzy C-regression models: Introduction to epsilon-insensitive fuzzy modeling [J]. IEEE Transactions on Systems, Man and Cybernetics, Part B: Cybernetics, 2004, 34: 4-15

[14] LeskiJ M. TSK fuzzy modeling based on ɛ-Insensitive learning [J]. IEEE Transactions on Fuzzy Systems, 2005, 13: 181-193

[15] ChuangC C, JengJ T, TaoC W. Hybrid robust approach for TSK fuzzy modeling with outliers [J]. Expert Systems with Applications, 2009, 36: 8925-8931

[16] SerneelsS, CrouxC, FilzmoserP, EspenP J V. Partial robust M-regression [J]. Chemometrics and Intelligent Laboratory Systems, 2005, 79: 55-64

[17] JiaR D, MaoZ Z, ChangY Q, ZhangS N. Kernel partial robust M-regression as a flexible robust nonlinear modeling technique [J]. Chemometrics and Intelligent Laboratory Systems, 2010, 100: 91-98

[18] ChiuS L. Fuzzy model identification based on cluster estimation [J]. Journal of Intelligent and Fuzzy Systems, 1994, 2: 267-278

[19] ChiuS L. An efficient method for extracting fuzzy classification rules from high dimensional data [J]. Advanced computational intelligence, 1997, 1(1): 1-7

[20] SetnesM, HellendonH. Orthogonal transforms for ordering and reduction of fuzzy rules [C]. The Ninth IEEE International Conference on Fuzzy Systems, Fuzz IEEE 2000, 2000San Antonio, T XIEEE700-705

[21] ChuF, WangF-l, WangX-g, ZhangS-ning. Performance modeling of centrifugal compressor using kernel partial least squares[J]. Applied Thermal Engineering, 2012, 44: 90-99

[22] ChuF, WangF-l, WangX-g, ZhangS-ning. A hybrid artificial neural network-mechanistic model for centrifugal compressor [J]. Neural Computing and Application, 2014, 24: 1259-1268

[23] ChuF, WangF-l, WangX-g, ZhangS-ning. A model for parameter estimation of multistage centrifugal compressor and compressor performance analysis using genetic algorithm [J]. Science China: Technological Sciences, 2012, 55(11): 3163-3175

[24] ChuangC C, JengJ T, LinP T. Annealing robust radial basis function networks for function approximation with outliers [J]. Neurocomputing, 2004, 56: 123-139