基于掩码分段匹配的否定选择数据分类

Method of negative selection data classification based on mask piecewise matching

根据免疫否定选择原理,设计了基于掩码分段匹配的否定选择分类器,用于实现规则匹配分类。给出了适用于免疫优化的分类规则编码及分类信息分的评价标准,通过免疫进化对其进行群体优化以生成更为简洁、便于理解的数据规则集。该方法使得免疫优化的各种优良特性在数据分类中得到充分的运用,避免了传统分类算法缺乏全局优化能力的缺点,提高了对样本的识别能力。实验结果表明,这种免疫分类器及优化方法是一种有效、可行的分类器设计方案,提高了数据分类的准确性。

Based on immune negative selection principle,a novel mask piecewise matching negative selection classifier is introduced to accomplish the date classification.In order to improve the accuracy of classified mining,a classification rule coding is defined and a criterion of information grade is proposed to obtain the rules.This method fulls use of all kinds of excellent traits of immune optimum in data classification and has a better command of obtain the global optimum than traditional algorithm.The simulation results illustrate that the immune classifier is an available and feasible algorithm for data classification,and improves the classification＇s accuracy and validity.