P-C-R模型构建及应用

针对目前预测模型精度低的问题，提出将主成分分析、聚类分析用于RBF神经网络预报建模，从而克服大样本数据提取的困难，使得指标的选取能更全面地反映状况，有效地缩减RBF网络的输入节点数并提高模型的预报精度。利用MATLAB的神经网络工具箱，实现了神经网络训练和仿真验证。仿真结果表明，该模型有较高的预报能力。提出的基于主成分分析和聚类分析的RBF网络预报模型---PCR模型为研究预报提供了一个新的思路和方法，并为其他领域的建模研究开阔了思路，具有一定的理论价值和的应用价值。

Application and Research of PCR model

In view of the present prediction problem of low precision,it puts forward the principal component analy-sis,clustering analysis are used in the RBF neural network prediction model,which can overcome the difficulties of large sample data extraction,makes the selection of indicators can more fully reflect the status,effectively reduce the number of input nodes of RBF neural network and improve the forecasting precision of model. Using the MATLAB neural network toolbox,realized the neural network training and simulation. The simulation results show that the model has higher prediction ability. The RBF neural network prediction model based on principal components analy-sis and clustering analysis-PCR model,which provides a new idea and method for the research of cement clinker strength prediction,open the way for the study of modeling of other fields. It has the certain theory value and appli-cation value.