基于改进的Elman神经网络的网络舆情预警模型  

Early-warning of Network Public Opinion Model Based on Modified Elman Neural Network

作　　者：毕殿杰[1] 张子振[1] 余永红[1] 张林[1] 

机构地区：[1]安徽财经大学管理科学与工程学院,安徽蚌埠233030

出　　处：《河北科技师范学院学报》2016年第3期10-15,共6页Journal of Hebei Normal University of Science & Technology

基　　金：安徽财经大学校级科学研究项目(项目编号:ACKY1651);安徽省高校自然科学研究重点项目(项目编号:KY2014A005)

摘　　要：在深入分析影响网络舆情发展变化的各种因素的基础上,建立了三层网络舆情指标体系。针对表征舆情事件样本数据的非线性动态变化等特点,提出一种改进的Elman神经网络模型。该模型以量化的网络舆情指标为网络输入,以网络舆情预警等级为输出。实验结果表明,该模型在一定程度上降低了预测问题的复杂度,同时又在一定程度上提高了预警精度,为科学分析网络舆情预警机制提供一种解决问题的思路。A three-class index system was constructed in terms of a detailed analysis of the various factors that might affect the development of the network public opinion. Furthermore, a modified Elman neural network model was proposed according to the nonlinear dynamic change of the sample data of public opinion events. The model characterized the network public opinion indexes as the input, and its warning level as the output. The results showed that the model not only reduced the forecast complexity but also improved the accuracy of early warning. The new model might provide a thought to solve the problems scientifically in the early-warning mechanism of network public oninion.

关 键 词：网络舆情 指标体系 ELMAN神经网络 预警模型 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程]