人工免疫记忆分类器类别不均匀分布分类研究

Research of Artificial Immune Memory Classifier in Non- skewed Class Distribution

类别分配不均匀是实际中常见的分类问题。文章利用基于免疫记忆的分类器—免疫记忆分类器研究该类问题。通过改进人工免疫记忆分类器距离度量方式,发现在比例选择平均距离度量情况下,该种分类器可以很好地解决类别分配不均匀问题。与另一种免疫分类方法AIRS和传统的KNN分类结果比较表明,人工免疫记忆分类器能够解决这类问题,效果好于后两者,为解决该类问题提供了新的思路和方法。

The non-skewed distribution of class is very common in real world. In this paper,artificial immune memory classifier is used to research such problem.By the ways of improving distance measure of artificial immune memory classifier,it is found that this kind of classifier can solve such problem well under the condition of proportion selection average distance measure.It is compared with another way based on artificial immune system-AIRS and traditional KNN.It shows that artificial immune memory classifier can solve such problem and has better performance than the two ways.It presents a new way for solving such problem.