一种自适应蚁群分类算法  

An Adaptive Ant Colony Classification Algorithm

作　　者：马安香[1] 张长胜[1] 张斌[1] 张晓红[1] 

机构地区：[1]东北大学信息科学与工程学院,辽宁沈阳110819

出　　处：《东北大学学报（自然科学版）》2014年第8期1102-1106,共5页Journal of Northeastern University(Natural Science)

基　　金：国家自然科学基金资助项目(61300019);中央高校基本科研业务费专项资金资助项目(N120404013);沈阳市科技专项基金资助项目(F11-264-1-35)

摘　　要：将分类学习看作是一个找出最优分类规则的优化问题,提出一种自适应蚁群分类算法——Adaptive L_AMP,以得到一组可理解的分类规则.在基于规则的分类方法中,规则评价函数的选取至关重要,本文提出的算法能够针对不同数据集自动选取与之相适应的规则评价函数以提高分类准确性.此外,为进一步提高算法的分类准确率,设计了一种局部搜索策略并将其融入到Adaptive L_AMP算法中.最后对算法进行了分析,并在多个公用的真实数据集上与相关算法进行了比较,结果表明Adaptive L_AMP算法能够更加有效地解决分类问题.Taking classification learning as an optimization problem of seeking for optimal classification rules, an adaptive ant colony algorithm, adaptive LAMP, is proposed to solve classification problem and get the comprehensible classification rules. In the rule-based classification methods, the way of choosing of a rule-evaluating function is important. The adaptive L_ AMP proposed can be used to automatically select an appropriate rule-evaluating function according to the data set, thus improving the classification correctness. Moreover, a local search technique is introduced into the algorithm set and compared with other relevant algorithms. proposed. The algorithm is run on the real data The results show the superiority of the algorithm proposed.

关 键 词：蚁群算法 自适应蚁群算法 分类 规则评价函数 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]