基于灰色神经网络的Web社区主题舆情传播模型

Web舆情传播的动态性，不确定性等特征给精确预测舆情传播带来困难．在分析了灰色理论系统的基础上，提出了灰色理论微分方程型模型（GM）和扩展BP神经网络的组合模型，该组合模型综合考虑了网络的结构和传播特性，首先建立灰色理论微分方程型模型，然后映射到扩展的BP神经网络中，通过训练数据来训练该神经网络，使网络具有传播预测能力．仿真实验表明，该组合模型在Web社区主题舆情传播预测精确性方面高于单一的GM模型．

Model of Web Community Network Public Opinion Spread Based on Gray Neural Network

In order to solve the propagation problem in the public opinion's spread process in the dynamic and uncertain Web community, we proposal a model which combines the gray theory differential equation model with the extended BP neural network model based on the analysis of gray theory system. Firstly, according to the network structure and propagation characteristics, the differential equation model of gray theory is established. Then, the model was mapped to the extended BP neural network. Finally, through training the neural network, the network model owns an ability of propagating prediction. Simulation results show that the combined model has a higher accuracy ratio than the single GM model in the propagation prediction of the Web community of public opinion's spread.