基于Rough集和构造性学习神经网络的经济预警模型  

Economic early-warning model based on rough sets and constructive learning neural networks

作　　者：朱勇[1] 吴涛[2] 

机构地区：[1]池州学院数学与计算机科学系,安徽池州247100 [2]安徽大学数学与计算科学学院,安徽合肥230039

出　　处：《合肥工业大学学报（自然科学版）》2007年第7期836-839,843,共5页Journal of Hefei University of Technology：Natural Science

摘　　要：在深入分析人工神经网络(ANN)与粗集(Rough set)理论方法的基础上,将Rough集、构造性神经网络与宏观经济预警研究相结合,尝试建立起基于Rough集-覆盖算法的构造性神经网络宏观经济预警方法体系;结合安徽省经济数据,对该模型预警实证分析。理论和实践证明,基于Rough集-覆盖算法的构造性神经网络预警模型是有效的、可行的,且具有较高的精度,从而为动态经济预警提供一条新的途经。Based on the introduction of the method of constructive learning artificial neural networks （ANN） and the theory of rough sets, the model of the macroeconomic early-warning system is built which is based on the rough set and the constructive learning ANN based on the covering algorithm. The economic data of Anhui Province from 1997 to 2004 are used to carry out an empirical study, which proves the feasibility and validity of the model. The proposed model has high precision, and it provides a new way for the study of dynamic economic early-warning.

关 键 词：经济预警 构造性学习 覆盖算法 粗糙集 

分 类 号：O159[理学—数学] TP183[理学—基础数学]