基于模糊关联规则的预测系统遗传优化*

作者：李燕,王锋 日期：2010-06-28/span> 浏览：3923 查看PDF文档

基于模糊关联规则的预测系统遗传优化*
李燕,王锋
（杭州电子科技大学 计算机学院,浙江 杭州 310018）

摘要：为提高预测系统中的预测精度，提出了一种基于模糊关联规则的优化的预测系统设计方法。该方法通过两个阶段来实现：首先采用竞争聚集算法得到各数量型属性优化的模糊集个数，从而挖掘出优化的模糊关联规则。在得到用于构建预测系统规则库的模糊关联规则后，采用遗传算法约简冗余规则库，实现精确性和解释性的折衷，以提高预测精度。最后将此方法运用于Abalone样本数据集进行实验分析，证实此方法解决了模糊关联规则的冗余问题，有效提高了预测精度。
关键词：模糊关联规则;预测系统;竞争聚集算法;遗传算法
中图分类号：TP18文献标识码：A文章编号：1001-4551(2010)06-0108-05

Optimization of fuzzy association rule based prediction
system by genetic strategies

LI Yan, WANG Feng
(College of Computer, Hangzhou Dianzi University, Hangzhou 310018, China)

Abstract: Aiming at improving the accuracy of prediction system, an optimized method which based on fuzzy association rules was proposed for prediction system designing. The method was constructed by two phase, competitive agglomeration algorithm was employed to partition quantitative attributes from each data record into several optimized fuzzy sets, resulting in an initial prediction system. And a genetic algorithm was employed to optimize rule base, which can achieve a trade-off between accuracy and interpretability. This approach was applied to the Abalone data set, and demonstrated that this method solved the redundancy in fuzzy association rules and effectively improved the prediction accuracy.
Key words: fuzzy association rules; prediction system; competitive agglomeration algorithm; genetic algorithm