粗糙集神经网络算法在数据挖掘中的研究与应用  

Research of Rough Set Neural Network on DM System

作　　者：王晓洁[1] 王付强[1] 

机构地区：[1]新乡师范高等专科学校计算机科学系,河南新乡453000

出　　处：《河南机电高等专科学校学报》2007年第4期23-25,共3页Journal of Henan Mechanical and Electrical Engineering College

摘　　要：神经网络是数据挖掘中最为常用的算法之一。它具有正确率高、抗噪声数据能力强、计算错误率低等优点。但神经网络算法也存在结构相对复杂、训练时间长、计算结果的可解释度比较低等问题。文中采用粗糙集理论对数据进行预处理,使用神经规则进行数据挖掘的新方法,该方法可以在结果精度有限降低的前提下,得到表示简单明确且错误率低的关联规则,同时可以减少网络训练时间,大大改进单独采用神经网络算法给系统带来的缺陷。Neural network is one of the most commonly used algorithms in data mining. It has the advantages of a high accuracy, a great anti-yawp capacity, and a low error rate in calculation. However, neural network also has its disadvantage that it has a relatively complicated structure, a longer training period and a deficiency in calculation report interpretation. This paper proposed a new method in which a preprocessing of data using rough sets theory will be followed by data mining using neural rules. With a limited decrease in accuracy, this method provides us with clear, definite and low error rate associate rules, as well as a shorter network training time. The proposed method thus eliminates some of the system limitations caused by of an exclusive/sole application of neural network.

关 键 词：粗糙集 神经网络 数据挖掘 关联规则 

分 类 号：TP274[自动化与计算机技术—检测技术与自动化装置]