基于FNN与GA相融合的数据挖掘方法研究  

Data Mining Study Based on Fuzzy Neural Networks and Genetic Algorithms

作　　者：王刚[1] 王本年[1] 

机构地区：[1]铜陵学院计算机系,安徽铜陵244000

出　　处：《计算机技术与发展》2008年第2期119-121,125,共4页Computer Technology and Development

基　　金：安徽省自然科学研究项目(KJ2007B382ZC)

摘　　要：模糊神经网络即具有输入信号是模糊量的神经网络,是模糊系统与神经网络相结合的产物,汇聚了二者的优点;遗传算法是一种自适应全局优化概率搜索算法。研究了基于模糊神经网络与遗传算法相融合的一种算法,在应用模糊神经网络进行数据挖掘前,应用遗传算法完成隶属函数的训练,以便更好地进行模糊神经网络学习;经过模糊神经网络学习后,提取相关规则,再次应用遗传算法,进行规则剪枝,提高数据挖掘效率。实验表明,与传统方法相比,该方法能够更快速、更加准确地进行数据挖掘,提取更精确的推理规则。With fuzzy input signals, fuzzy neural network is a combination of fuzzy system and neural network. Therefore, it asemhles the advantages of them. Genetic algorithm can adapt to the overall situation and better probability seeking automatically. Based on the combination of fuzzy neural network and genetic algorithm, studies an algorithm that can train the membership function by applying genetic algorithm before data mining by applying fuzzy neural network. Rules can be obtained after the study of fuzzy neural network while pruned after the application of genetic algorithm; thus the efficiency of data mining is increased. Experiments have shown that, compared with conventional algorithm, this one can mine data more quickly and more accurately, and get the reasoing rules more precisely.

关 键 词：数据挖掘 模糊神经网络 遗传算法 隶属函数 规则剪枝 

分 类 号：TP311[自动化与计算机技术—计算机软件与理论]