基于蚁群优化算法的基因选择*

将蚁群优化算法(ant colony optimization algorithm，ACO)引入基因选择领域，并用基因与类别的相关性分析所得值来初始化最优化问题，缩短了找寻最优解的时间；以基因子集整体的样本辨别能力与子集中基因之间的平均距离的线性表达作为目标函数，有利于在找到关键基因的同时消除冗余；同时，由于目标函数不采用分类准确度，大大降低了计算复杂度，提高了方法的灵活性和适应性。

Gene selection based on ACO algorithm

This paper carried ant colony optimization(ACO) algorithm into the domain of gene selection,and used the relati-vity between genes and sample classes of initialize the problem to reduce the runtime.In addition,it regarded the linear expression of the ability of distinguishing the samples of genes and the average distance among genes in the gene subset as the target function,which was in favor of finding the key genes and avoiding redundancy.At the same time,computing the target function has a low complexity because of its own structure,and the whole method has better agility and adaptability.