The Dynamic Fuzzy Measure For Multi-classifier Fusion

As the artificial intelligence science develop,new classifiers appear constantly. But in many problems only one classifier cannot meet our requirement. The multi-classifier fusion system offer us a way which can improve the performance of the classification system.Fuzzy Integral as an aggregation tool in multi-classifier fusion has attracts much attention.This method can improve the accuracy of classification and the robustness of systems.In multi-classifier fusion based on fuzzy integrals,fuzzy measures is one of key factors.If we choose appropriate fuzzy measures,the accuracy of classification can be improved distinctly. In this thesis we analyze what influence fuzzy measures have on the classification of the fusion system with given classifiers.It is discovered that multi-classifier fusion systems based on fuzzy integrals have some ability to correct the misclassification of classifiers.It is possible that the fusion system classifies correctly even if all classifiers misclassified one sample.And we use the neural networks to determine the dynamic fuzzy measures such that fuzzy measures can variety according to input sample to reflect the importance of each classifier and the interaction among them.The proposed method is feasible from the experiments.