基于遗传算法和模糊积分的多分类器集成

多分类器联合是解决复杂模式识别问题的有效办法。模糊积分是其中一种多分类器联合方法。但是对于模糊积分。如何计算模糊积分密度是一个尚未解决的问题。本文提出了一种基于模糊积分和遗传算法的分类器集成方法，该方法利用遗传算法计算模糊积分密度函数，再利用模糊积分把分类器输出信息联合起来。实验结果表明，该方法比其他方法能够得到更好的识别性能。

A COMBINATION METHOD OF MULTIPLE CLASSIFIERS BASED ON FUZZY INTEGRAL AND GENETIC ALGORITHMS

Combination of multiple classifiers is a effective solution to the problem of complicated pattern recognition.Fuzzy integral is one of the ways to combine classifiers,but how to calculate the fuzzy integral density is a problem which remains unsolved.A combination method based on fuzzy integral and genetic algorithms,which gets fuzzy integral density function with genetic algorithms and combines the output information of all classifiers with fuzzy integral,is presented.It has been shown by experiments that this method could achieve a more promising performance than other methods.