A Classifier Ensemble Method for Fuzzy Classifiers
In this paper, a classifier ensemble method based on fuzzy integral for fuzzy classifiers is proposed. The object of this method is to reduce subjective factor in building a fuzzy classifier, and to improve the classification recognition rate and stability for classification system. For this object, a method of determining fuzzy integral density based on membership matrix is proposed, and the classifier ensemble algorithm based on fuzzy integral is introduced. The method of selecting classifier sets is also presented. The proposed method is evaluated by the comparison of experiments with standard data sets and the existed classifier ensemble methods.
KeywordsRecognition Rate Individual Classifier Training Pattern Classifier Ensemble Fuzzy Measure
Unable to display preview. Download preview PDF.
- 2.Cordon, O., del Jesus, M.J., Herrera, F.: A proposal on reasoning methods in fuzzy rule-based classification systems. Int. J. of Approximate Reasoning 20(1), 21–45 (1999)Google Scholar
- 3.Wierzechon, S.T.: On fuzzy measure and fuzzy integral. In: Fuzzy information and decision processes, pp. 78–86. North-Holland, New York (1982)Google Scholar
- 4.Schapire, R.E.: The strength of weak learn ability. Machine Learning 5(2), 197–227 (1990)Google Scholar
- 6.Turner, K., Gosh, J.: Error correlation and error reduction in ensemble classifiers. Dept. of ECE, University of Texas, Texas (1996)Google Scholar
- 9.Blake, C.L., Merz, C.J.: UCI Repository of machine learning databases (1998), http://www.ics.uci.edu/~mlearn/MLRepository.html
- 10.Zhi-Hua, Z., Shi-Fu, C.: Neural Network Ensemble. Chinese Journal of Computers 25(1), 1–8 (2005)Google Scholar
- 11.Xie-Dong, Z.: Fuzzy Information Managing and Application, pp. 174–176. Science Publishing Company, Beijing (2003)Google Scholar
- 12.Ai-Min, Y.: The Model Research on Fuzzy Classification. Doctor’s Degree Paper of Fudan University (2005)Google Scholar