基于模糊神经网络的粗糙集在股市预测中的应用  

Rough Set Based on Fuzzy Neural Network for the Application in Prediction of Stock Market

作　　者：叶德谦[1] 马志强[1] 李帼[1] 姜皇普[1] 

机构地区：[1]燕山大学中德信息技术合作研究所,秦皇岛066004

出　　处：《计算机科学》2008年第4期168-169,183,共3页Computer Science

摘　　要：提出在模糊神经网络中使用粗糙集理论进行网络的设计。在模糊神经网络中引入粗糙集理论,不仅可以去除模糊神经网络中输入层的冗余神经元而且可以确定隐含层神经元的数目,从而使模糊神经网络具有更准确的逼近收敛能力和较高的精度。最后应用于股票市场,在股票买卖时机预测中取得了良好的效果。A new scheme of knowledge encoding in a fuzzy neural networks using rough set theoretical concepts is described. Introducing the rough theory in the fuzzy neural network can not only remove redundant neurons in the input layers of fuzzy neural, but also construct the neurons in the hidden layers. Thus enable the fuzzy neural network to have more accurate restraining ability and the high precision. Finally applies in the stock market, the stock business opportunity forecast has obtained the good effect.

关 键 词：粗糙集 模糊聚类 神经网络 股市预测 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程] F832.5[自动化与计算机技术—控制科学与工程]