Title :
Combination of Multiple K-NNCS by Fuzzy Integral
Author :
Wang, Li-juan ; Wang, Xiao-long ; Chen, Qing-cai
Author_Institution :
Dept. of Comput. Sci. & Technol., Harbin Inst. of Technol.
Abstract :
The k-NNC error varies with the value of k, when the number of training patterns is fixed. Therefore the combination of multiple k-NNCs with different k may improve the accuracy of the whole system. In this paper, multiple k-NNCs with different k are combined by fuzzy integral. The value of k and the density value for each classifier are automatically learned by GA. Experiments on three UCI databases have shown that the combined classification accuracy outperforms that of the single NNC
Keywords :
fuzzy set theory; genetic algorithms; learning (artificial intelligence); pattern classification; GA; UCI database; fuzzy integral; fuzzy measure; genetic algorithm; multiple-classifier combination; nearest neighbor classifier; training pattern; Classification algorithms; Computer errors; Computer science; Cybernetics; Databases; Decision making; Euclidean distance; Machine learning; Mathematics; Nearest neighbor searches; Voting; Fuzzy integral; Fuzzy measure; GA; Nearest neighbor classifier (NNC); multiple-classifier combination;
Conference_Titel :
Machine Learning and Cybernetics, 2006 International Conference on
Conference_Location :
Dalian, China
Print_ISBN :
1-4244-0061-9
DOI :
10.1109/ICMLC.2006.258979
