摘要
目前的模型选择方法对用户的期望过高 ,要求用户对DSS模型库中的模型和所要解决的问题都要有深刻的理解 ,这就导致了模型选择的实用性较差 .提出了一种基于专家系统和人工神经网络的模型自动选择理论和方法 ,这种理论和方法将模型选择分为模型类型选择和模型结构选择两部分 ,应用人工智能技术完成模型类型的自动选择 ,应用神经网络技术实现模型结构的自动选择 .特别是在模型结构的自动选择中 ,人工神经网络方法的工作原理主要由两个阶段组成 ,即学习和选择阶段 .在学习阶段 ,对每一模型类的模型组训练一个神经网络 ,使训练好的给定类模型的神经网络能够对用户给出一组数据的自动选择 (判断 )出适当的模型结构 .而且 ,这个方法通过趋势外推型模型类的实验 ,证明是很有效的 .
Model selection is an important component of model management system of decision support systems(DSS). A crucial problem confronting users of DSS is that the present methods of model selection require the users too much. The user are required to have a profound understanding of the models in the model base and the problems to be solved. This leads to poor feasibility of model selection. Presents a good combination of theory with method for automatic selection of model based on integration of expert system with artificial neural networks which is divided into, model class selection and model structure selection the use of an expert system technology to select model class and the use of an artificial neural network for automatic selection of model structure the artificial neural network method consisting of learning and working in particular and concludes from experimental results that what is proposed is an effective approach.
出处
《哈尔滨工业大学学报》
EI
CAS
CSCD
北大核心
2001年第1期24-27,123,共5页
Journal of Harbin Institute of Technology
基金
国家自然科学基金资助项目 !(795 0 0 0 0 3 )
