Fuzzy Approach to Combining Parallel Experts Response
This paper gives description of a parallel neural expert system for solving complex diagnostic problems. The idea of this paper is to propose an ensemble composed of independently trained neural networks. The aim is to build a classifier, which does not need a heavy training process and has good generalisation properties. This work contains a comparison of a few basic methods for combining multiple classifiers. Two classification problems are considered: breast cancer diagnosis and hand-written digit recognition.
KeywordsMultiple Classifier Fuzzy Measure Plurality Rule Trained Neural Network Neural Network Ensemble
Unable to display preview. Download preview PDF.
- 2.Korbicz, J., Uciski, D., Obuchowicz, A.: Artificial Neural Networks. Foundamentals and Applications, AOW, Warszawa (1994) (in Polish)Google Scholar
- 3.Mangasarian, O., Street, N., Wolberg, W.: Breast Cancer Diagnosis and Prognosis via Linear Programming, Mathematical Programming Technical Report 94–10 (1994) (available via WWW from ftp.cs.wisc.edu/math-prog)Google Scholar
- 5.Sung-Bae, Cho: Neural-network classifiers for recognizing totally unconstrained handwritten numerals, IEEE Trans. on Neural Networks, Vol. 8, No. 1 (1997) 4353Google Scholar
- 6.Xu, L., Krzyzak, A., Suen, Ch.: Methods of Combining Multiple Classifiers and Their Applications to Handwriting Recognition, IEEE Trans. on Systems, Man, Cybernetics, Vol.22,. No. 3 (1992) 418–435Google Scholar