应用分类器组合模型进行岩性识别

为了达到应用模式识别对储层岩性进行准确的识别,根据不同性质分类器的组合可以较全面描述一个模式从而降低识别错误及增强识别鲁棒性这种特性,结合模糊数学中的隶属度实现了多个分类器的定量组合.应用了几种不同的分类器进行组合,同时应用某油田的岩性识别作为实例进行验证,并与单个分类器进行比较,结果表明在间等条件下组合分类器的性能要比单个分类器的性能要好得多.这说明了应用不同性质的分类器进行组合识别的分类方法足可行性和有效性.

Application of Classifiers Combination in the Lithology Recognition

In order to use pattern recognition to accurately identify the reservoir lithology,the paper based on the combination of different classifiers can comprehensively describe a pattern recognition to reduce errors and improve the robustness of the identification features,and combined with Fuzzy membership in the realization of a number of quantitative composition of the classifiers.The results of experiment and comparison show that the method is feasible and effective.