摘要
将粗糙集理论（RST）与模糊神经网络（FNN）相结合，提出了一种基于粗糙集理论的模糊神经网络（RST-FNN）模型。新模型利用粗糙集的知识约简对样本数据去噪消冗，提取最优规则，从而克服模糊神经网络的“维数爆炸”灾难。实例仿真的结果表明，该模型的预测准确性较高，且具有结构精简、收敛速度快及泛化能力强等特点。
A rough fuzzy neural network （RST-FNN） model was proposed by combined the rough set theory （RST） with fuzzy neural networks （FNN）. The new model can overcome the curse of dimensionality by using the reduction of knowledge based on rough set theory to eliminate redundant and noise of the sample data. The simulation result indicates that the predictive accura-cy of the model is high. Moreover, it has characteristics of simple structure, fast convergence speed, and stronger generalization ability.
作者
黄福员
HUANG Fu-yuan （Business School of Zhanjiang Normal College, Zhanjiang 524048,China）
出处
《电脑知识与技术》
2013年第11期7078-7082,7095,共6页
Computer Knowledge and Technology
基金
该文受广东省自然科学基金项目（10452404801006352）、广东高校优秀青年创新人才培育项目（WYM10103）及湛江师范学院项目（W0817）资助
关键词
金融风险预警
模糊神经网络
粗糙集理论
Financial Risk Early Warning
Fuzzy Neural Networks
Rough Set Theory
