ARTIFICIAL IMMUNE SYSTEM CLASSIFICATION OF MULTIPLE- CLASS PROBLEMS

A new classifier, AIRS, based on the principles of resource-limited artificial immune systems, has been shown recently to consistently rank among the best five to eight known classifiers for a number of well-studied classification problems, including the Iris data, the Cleveland heart disease data, and others. However, each of these previous test problems has involved only a few classes. In this paper, we discuss the general application of AIRS to multi-class problems, and we compare it to a similar well-known classifier, Kohonen's LVQ, on both simulated and real-world data sets.

[1]  Alan S. Perelson,et al.  Self-nonself discrimination in a computer , 1994, Proceedings of 1994 IEEE Computer Society Symposium on Research in Security and Privacy.

[2]  Catherine Blake,et al.  UCI Repository of machine learning databases , 1998 .

[3]  A. B. Watkins,et al.  A resource limited artificial immune classifier , 2002, Proceedings of the 2002 Congress on Evolutionary Computation. CEC'02 (Cat. No.02TH8600).

[4]  A. B. Watkins,et al.  A new classifier based on resource limited artificial immune systems , 2002, Proceedings of the 2002 Congress on Evolutionary Computation. CEC'02 (Cat. No.02TH8600).

[5]  Jonathan Timmis,et al.  A resource limited artificial immune system for data analysis , 2001, Knowl. Based Syst..

[6]  Thomas Richardson,et al.  Interpretable Boosted Naïve Bayes Classification , 1998, KDD.

[7]  Geoffrey I. Webb.  Further Experimental Evidence against the Utility of Occam's Razor , 1996, J. Artif. Intell. Res..

[8]  Fernando José Von Zuben,et al.  Learning and optimization using the clonal selection principle , 2002, IEEE Trans. Evol. Comput..