基于云模型的网络舆情预警方法

A Method for Early Warning of Internet Public Opinions Based on Cloud Model

网络舆情预警是网络舆情研究的主要目的。本文从3个方面改进现有网络舆情预警方法：首先，制定客观有效的预警体系；其次，通过跟踪统计特性、构建备选模型库、制定模型选择规则等3个步骤，来自适应地预测网络舆情的发展过程；第三，运用云模型实现从预警指标到威胁状态的转换，并将当前威胁状态与潜在威胁趋势进行融合评估，得到预警等级。实验表明，该方法得到的预警等级具有时变性，变化过程与事件发展过程一致，并体现了网民的上网规律，符合客观实际。

Early warning of Internet public opinions is the main purpose of the research. This paper improves the existing early warning methods of Internet public opinions from three aspects. Firstly, this paper establishes objective effective early warning system. Secondly, by tracking the statistical characteristics, constructing alternative model library and making model selection rules, this paper predicts the development process of Internet public opinions adaptively. Thirdly, this paper converts the early warning indexes into the threat state by cloud model, and gets the early warning level after fusing the current threat state and the potential threat trend. The experimental results show that the early warning level obtained by this method is time-varying and that the change process of the early warning level is consistent with the event development process, reflects the netizen＇s pattern of surfing the Interuet and accords with the objective reality.