从数据库中提取规则的神经网络方法

Neural networks methods of extracting rule from database

评述了利用神经网络从数据库中进行规则发现的几种方法 ,采用权值组合算法提取规则 ;利用模糊推理神经网络 ,采用CamDelta算法提取模糊规则 ;基于从数据中提取模糊控制规则 ;利用生长自组织映射神经网络 ,采用分级聚类SOM算法发现规则 ;利用CFNet网络 ,基于可信度因子 ,提取不确定性规则 ;利用模糊颗粒神经网络 ,采用启发式学习算法 ,从数值 -语言数据中发现规则 提出了数据库中提取规则所面临的几个问题 ,以及解决这些问题的某些思路 具体提出了一种分布式环境下基于多Agent技术的规则提取方法 图 6,参

Several methods of extracting rules from database using neural networks are surveyed in this paper:extracting rules based on BP network and BP algorithm using combining weight algorithm.Extracting fuzzy rules using fuzzy inferencing neural networks and CamDelta algorithm.Extracting fuzzy control rules from data based on the competitive neural networks and DCL algorithm.Discovering rules using the growing self-organizing map neural networks and hierarchical clustering algorithm.Extracting noncertainty rules using CFNet networks with certainty factor.Discovering rules numerical-linguictic data using fuzzy granular neural networks and heuristic learning algorithm.Finally,some confronting problems for extracting rules and ideas to solve are presented.And a rule extracting methods based on multiagent in distributed environment is contructed.6figs.,17refs.