基于信息集成技术的网络舆情预警方法研究

The public opinion has become a barometer of varieties of social phenomena and problems. Online public opinions pre-warning, as a hot topic of online public opinions researches, has been increasingly pain attentions. Recently, information aggregation technologies develop rapidly, and have been successfully applied in many fields, but methods of online public opinions pre-warning based on information aggregation techniques are seldom studied. The online public opinions pre-warning problems are typical multi-criteria decision making ones. We here focus on online public opinions pre-warning methods based on information aggregation technologies: including information measurements and information integration technologies, so as to lay the theoretical foundation for methods of online public opinions pre-warning; research online public opinions pre-warning index design, based on the indicators described in the language, using the theory of word computing and dual hesitant fuzzy sets, more objective and more reasonable to reflect the actual situation of network public opinion, and to facilitate the numerical analysis, and study the method to determine index weight; Based on the above methods, we will research outranking methods (such as ELECTRE, PROMETHEE etc.) and power averaging operators of multi-criteria decision making, so as to design decision making methods for online public opinions pre-warning. We will provide new methods from two perspectives of information representations and pre-warning decision making, which is strongly practical. It is meaningful and valuable on theories and applications for promoting the ability of online public opinions pre-warning and providing decision aids.

网络舆情是反映社会运行态势的“晴雨表”，网络舆情预警作为网络舆情研究的热点之一越发受到重视。近年来，信息集成技术发展迅速，但基于信息集成技术的网络舆情预警方法还缺少深入的研究。网络舆情预警是一个典型的多属性决策问题。本项目研究基于信息集成技术的网络舆情预警方法：研究信息测度和信息集成技术，为网络舆情预警方法奠定理论基础；研究相应的预警指标设计方法，在语言指标描述的基础上，采用对偶犹豫模糊集和词计算等理论量化成数值指标，更客观更合理的反映网络舆情实际情况，同时便于数值分析，并研究指标权重确定方法；基于本项目研究的预警指标和权重确定方法，研究多属性决策的ELECTRE、PROMETHEE的Outranking法和幂均算子方法，提出网络舆情预警方法。本项目可从信息表示和预警决策两个方面为网络舆情预警提供新方法，可操作性强，在提升网络舆情预警能力、提供决策支持等方面具有较大的理论意义和应用价值。

我国正处于改革发展的攻坚期和社会转型期，网络舆情是反映社会运行态势的“晴雨表”，网络舆情预警作为网络舆情研究的热点之一越发受到重视。近年来，信息集成技术发展迅速，但基于信息集成技术的网络舆情预警方法还缺少深入的研究。网络舆情预警是一个典型的多属性决策问题。针对突发事件下舆情疏导面临的舆情信息多源异构、舆情传播动态多变、舆情疏导滞后差等问题，以提升舆情治理能力为目标，基于微观挖掘与宏观分析相结合、关联耦合与动态演化相结合、快速优化与智能决策相结合的思路。本项目研究基于信息集成技术的网络舆情预警方法：研究信息测度和信息集成技术，为网络舆情预警方法 奠定理论基础；研究相应的预警指标设计方法，在语言指标描述的基础上，采用对偶犹豫 模糊集和词计算等理论量化成数值指标，更客观更合理的反映网络舆情实际情况，同时便于数值分析，并研究指标权重确定方法；基于本项目研究的预警指标和权重确定方法，研究多属性决策的ELECTRE、PROMETHEE的Outranking法和幂均算子方法，提出网络舆情预警方法。本项目可从信息表示和预警决策两个方面为网络舆情预警提供新方法，实现多源异构舆情信息的深度集成、舆情传播的精准演化、舆情疏导的快速优化，以期形成完整、实用的网络舆情传播疏导应用示范。