摘要
提出了一种基于最大-最小蚁群算法的模糊分类系统设计方法。该方法通过两个阶段来实现:特征变量选择和模型参数优化。首先采用蚁群算法对特征变量进行选择,得到一组具有较高分辩性能的特征变量,提高模型的解释性;在模型结构确定后,蚁群算法从训练样本中提取信息对模型的参数进行优化,在保证模型精确性的前提下,构造具有较少变量数目及规则数目的模糊模型,实现了精确性与解释性的折衷。最后将本方法运用到Iris和Wine数据样本分类问题中,并将结果与其它方法进行比较,仿真结果证明了该方法的有效性。
An approach to design fuzzy classification systems based on the Max-Min ant colony algorithm is proposed. The method consists of two phases： a feature selection phase and a model parameter optimization phase. In order to obtain fuzzy models with a good interpretability-accuracy trade-off, the method employs ant colony algorithm to sort out a group of input variables from training patterns with high performance of identification. After the identification of the surface structure, ant colony algorithm is applied to identify the parameters of the Membership Functions used in the rule base, taking into account the information provided by training patterns. Max-Min Ant System （MMAS）, a good alternative to existing algorithms, is presented to optimize the fuzzy modeling. The performance of the proposed method both for training and test data is examined by computer simulations on the Iris and Wine data classification problems.
出处
《模糊系统与数学》
CSCD
北大核心
2008年第4期87-98,共12页
Fuzzy Systems and Mathematics
基金
国家自然科学基金资助项目(60674001)
863安全保障课题(2007AA11Z247)
关键词
蚁群算法
模糊分类系统
解释性
Ant Colony Algorithm
Fuzzy Classification System
Interpretability
