Abstract
Multiple nearest neighbor classifier system (MNNCS) is a popular method to relax the curse of dimensionality. In previous work, most of the MNNCSs are designed by random methods. Random methods may generate unstable component classifiers. In order to relax the randomness, large amount of component classifiers are needed. This paper first extends nearest neighbor classifier into fuzzy nearest neighbor classifier, and proposes a new multiple fuzzy nearest neighbor classifier system based on mutual information and fuzzy integral, called MIFI-MFNNCS. MIFI-MFNNCS adopts target perturbation. Target perturbation decomposes the original classification problem into several sub-problems, where one sub-problem represents one class data. Each sub-problem is described by the relevant data and features. Then it is classified by one component classifier. Therefore, the number of component classifiers can be fixed and reduced. For one component classifier, data may be selected according to its class. And feature is needed to be selected by mutual information. Mutual information can reduce the uncertainty of each component classifier. Feature selection by mutual information in MIFI-MFNNCS may be less affected by the interaction among different classes. The diversity decisions from sub-problem classifiers are combined by fuzzy integral to get the final decision. Here we propose a new method to compute density value according to mutual information, which is a simple method. To demonstrate the performance of the proposed MIFI-MFNNCS, we perform experimental comparisons using five UCI datasets. The results of component classifiers in MIFI-MFNNCS for Ionosphere are shown and analyzed. MIFI-MFNNCS is compared with (1) NNC (2) NNC after feature selection by mutual information (MI-FS-NNC). In multiple fuzzy nearest neighbor classifier system (MFNNCS), mutual information is compared with attribute bagging. And three combination methods are compared, including fuzzy integral, majority voting rule and average. The experimental results show that the accuracy of MIFI-MFNNCS is better than other methods. And mutual information is superior to attribute bagging. Fuzzy integral shows a better performance than majority voting rule and average.
Similar content being viewed by others
References
Ludmila I (2004) Kuncheva, combining pattern classifiers: methods and algorithms, wiles
Saranlm A, Demirekler M (2001) On output independence and complementariness in rank-based multiple classifer decision systems. Pattern Recogn 34:2319–2330
Zhou ZH (2005) Ensembling local learners through multimodal perturbation. IEEE Trans SMC Part B 35(4):725–735
Breiman L (1996) Bagging predictors. Mach Learn 24(2):123–140
Freund Y, Schapire RE (1995) A decision-theoretic generalization of online learning and an application to boosting. Proceedings of the 2nd European conference on computational learning theory, Barcelona, Spain, pp 23–37
Ho TK (1998) The random subspace method for constructing decision forests. IEEE Trans Pattern Anal Mach Intell 20(8):832–844
Brylla R, Gutierrez-Osunab R, Queka F (2003) Attribute bagging: improving accuracy of classifer ensembles by using random feature subsets. Pattern Recogn 36:1291–1302
Langley P, Iba W. Average-case analysis of a nearest neighbor algorithm. Proceedings of the thirteenth international joint
Bay SD (1998) Combining nearest neighbor classifiers through multiple feature subsets. In: Proceedings of the 15th international conference on machine learning, Madison, WI, pp 37–45
García-Pedrajas N, Ortiz-Boyer D (2009) Boosting k-nearest neighbor classifier by means of input space projection. Expert Syst Appl 36(7):10570–10582
Oza NC, Tumer K (2001) Input decimation ensembles: decorrelation through dimensio-nality reduction, 2nd international workshop on multiple classifier systems. In: Kittler J, Roli F (eds) Lecture notes in computer science. 2096:238–247
Chow TWS, Huang D (2005) Estimating optimal feature subsets using efficient estimation of high-dimensional mutual information. IEEE Trans Neural Netw 16(1):213–224
Batttiti R (1994) Using mutual information for selecting features in supervised neural net learning. IEEE Trans Neural Netw 5(4):537–550
Kwak N, Choi CH (2002) Input feature selection for classification problems. IEEE Trans Neural Netw 13(1):143–159
Kwak N, Choi CH (2002) Input feature selection by mutual information based on parzen window. IEEE Trans PAMI 24(12):1667–1671
Huang J, Cai Y, Xu X (2007) A hybrid genetic algorithm for feature selection wrapper based on mutual information. Pattern Recogn Lett 28(13):1825–1844
François D, Rossib F, Wertza V, Verleysen M (2007) Resampling methods for parameter-free and robust feature selection with mutual information. Neurocomputing 70(7–9):1276–1288
Liu HW, Sun JG, Liu L, Zhang HJ (2009) Feature selection with dynamic mutual information. Pattern Recogn 42(7):1330–1339
Wang ZY, Klir GJ (1992) Fuzzy measure theory. Plenum Press, New York
Keller JM, sborn OJ (1996) Training the fuzzy integral. J Approx Reason 15:1–24
Wang ZY, Leung KS, Wang J (1999) A genetic algorithm for determining nonadditive set functions in information fusion. Fuzzy Sets Syst 102(3):463–469
Wang J, Wang ZY (1997) Using neural networks to determine Sugeno measures by statics. Neural Netw 10(1):183–195
Hu BG, Wang Y (2008) Evaluation criteria based on mutual information for classifications including rejected class. Acta Automatica Sin 34(11):1396–1403
Wang LJ (2006) Combination of multiple K-NNCs by Fuzzy integral. Proceedings of the fifth international conference on machine learning and cybernetics, pp 1774–1778
Sungeno M (1997) Fuzzy measures and fuzzy integrals—a survey. In: Gupta MM, Saridis GN, Gaines BR (eds) Fuzzy automata and decision processes. North-Holland, Amsterdam, pp 89–102
Acknowledgments
This paper is supported by Doctor initial fund of Guangdong university of technology (093031), Key Basic Research Project of Hebei Province Applied basic research plan (08963522D), and Hebei Province Department of Education Planning Project (Z2008302). Some parts of this paper has been appeared in Multiple Nearest neighbor classifiers system based on feature perturbation by mutual information, which has published in Proc of ICMLC 10.
Author information
Authors and Affiliations
Corresponding author
Rights and permissions
About this article
Cite this article
Wang, L.J. An improved multiple fuzzy NNC system based on mutual information and fuzzy integral. Int. J. Mach. Learn. & Cyber. 2, 25–36 (2011). https://doi.org/10.1007/s13042-010-0006-8
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s13042-010-0006-8