Abstract
This paper presents a method to find a model of a system based on the integration of a set of local models. Mainly, properties are sought for the local models: independence of clusters and interpretability of their validity. This has been achieved through the introduction of a possibilistic clustering for the first property and a pre-fixed shape of the membership functions for the second one. A new cost index for the clustering optimization problem has been defined consisting of two terms: one for global error and another for local errors. By giving higher importance to the local errors term, local models valid regionally can be found. To avoid local optima and numerical issues, the parameters of the models are found using global optimization. This new method has been applied to several data sets, and results show how the desired characteristics can be achieved in the resulting models.
Similar content being viewed by others
References
Aronovich L, Splieger I (2010) Bulk construction of dynamic clustered metric trees. Knowl Inf Syst 22(2): 211–244
Barcelo-Rico F, Diez J, Bondia J (2010) A comparative study of codification techniques for clustering heart disease database. Biomed Signal Process Control. doi:10.1016/j.bspc.2010.07.004
Bezdek JC (1981) Pattern recognition with fuzzy objective functions algorithms. Plenum Press, New York
Bezdek JC, Pal NR (1998) Some new indexes of cluster validity. IEEE Trans Syst Man Cybern Part B Cybern 28: 301–315
Bezdek J, Ehrlich R, Full W (1984) Fcm: The fuzzy c-means clustering algorithm. Comput Geosci 10: 191–203
Cheng K, Liu L (2009) “best k”: critical clustering structures in categorical datasets. Knowl Inf Syst 20(1): 1–33
de Oliveira J, Pedrycz W (2007) Advances in fuzzy clustering and its applications. Wiley, New York
De Carlo LT (1997) On the meaning and use of Kurtosis. Psychol Methods 2: 292–307
Diez JL (2003) Técnicas de agrupamiento para identificacin y control por modelos locales. PhD thesis, Universitat Politècnica de València
Diez JL, Navarro JL, Sala A (2007) A fuzzy clustering algorithm enhancing local model interpretability. Soft Comput Fusion Found Method Appl 11: 973–983
Diez JL, Sala A, Navarro JL (2005) Target shape possibilistic clustering applied to local-model identification. Engineering Applications of Artificial Intelligence 4th
Kim EY, Kim SY, Ashlock D, Nam D (2009) Multi-k: accurate classification of microarray subtypes using ensemble k-means clustering. BMC Bioinform 10: 260
Egea JA, Rodriguez-Fernandez M, Banga JR, Mart R (2007) Scatter search for chemical and bio-process optimization. J Global Optim 37: 481–503
Egea-Larrosa JA (2008) New Heuristics for Global Optimization of Complex bioprocesses. PhD thesis, Universidade de Vigo
Gustafson EE, Kessel WC (1978) Fuzzy clustering with a fuzzy covariance matrix. In: IEEE conference on decision and control, pp 761–766
Hartigan J, Wong MA (1979) A K-means clustering algorithm. JR Stat Soc Ser C 28: 100–108
Hathaway R, Bezdek J (1993) Switching regression models and fuzzy clustering. IEEE Trans Fuzzy Syst 1(3): 195–204
Abonyi J, Babuska R, Szeifert F (2002) Modified gath-geva fuzzy clustering for identification of takagi-sugeno fuzzy models. IEEE Trans Syst Man Cybern Part B Cybern 32(5): 612–621
Krishnapuram R, Keller JM (1993) A possibilistic approach to clustering. IEEE Trans Fuzzy Syst 1: 98–110
Emami MR, Turksen IB, Goldenberg AA (1998) Development of a systematic methodology of fuzzy logic modeling. Trans Fuzzy Syst 6: 346–366
Goebel M, Gruenwald L (1999) A survey of data mining and knowledge discovery software tools. SIGKDD Explor Newsl 1(1): 20–33
Ryoke M, Nakamori Y, Suzuki K (1995) Adaptive fuzzy clustering and fuzzy prediction models. In: Fuzzy Systems, 1995. International Joint Conference of the Fourth IEEE International Conference on Fuzzy Systems and The Second International Fuzzy Engineering Symposium., Proceedings of 1995 IEEE International Conference on vol 4
Sugeno M, Yasukawa T (1993) A fuzzy-logic based approach to qualitative modelling. Trans Fuzzy Syst 1: 7–31
Chaoji V, Hasan MA, Salem S, Zaki M (2009) Sparcl: an effective and efficient algorithm for mining arbitrary shape-based clusters. Knowl Inf Syst 21(2):201–229
Author information
Authors and Affiliations
Corresponding author
Rights and permissions
About this article
Cite this article
Barceló-Rico, F., Díez, JL. & Bondia, J. New possibilistic method for discovering linear local behavior using hyper-Gaussian distributed membership function. Knowl Inf Syst 30, 377–403 (2012). https://doi.org/10.1007/s10115-011-0385-5
Received:
Revised:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s10115-011-0385-5