摘要
[目的/意义]互联网的开放式传播给网络舆情的监管和治理带来困难。准确地预测网络舆情事件能够帮助政府等相关部门及时、有针对性地采取引导措施,控制网络舆情的传播。[方法/过程]首先采集网络舆情数据构建事理图谱,通过改进聚类算法实现舆情事件泛化,构建抽象事理图谱。根据抽象事理图谱中事件演化方向和概率大小,预测网络舆情事件。[结果/结论]医疗网络舆情实证结果表明,该方法可以较好地预测舆情事件,准确率达到72.03%。网络舆情事件预测有效地补充了现有网络舆情预测研究仅关注热度、情感和话题的不足,为更精准地实现网络舆情治理提供了支持。
[purpose/significance]The network’s wide open communication will bring difficulties to the supervision and guidance of Internet public opinion.Accurately predicting network public opinion can help the government and other relevant departments to take timely and targeted measures to control the spread of internet public opinion.[Method/process]Firstly,the research collects internet public opinion data to construct the event evolution graph,and then use the improved clustering algorithm to generalize specific event to construct the abstract event evolution graph.Next,the paper predict event which happened next based on the evolution direction and weight.[Result/conclusion]The empirical example of medical internet public opinion shows that this method can better predict the public opinion events,and it’s accuracy is 72.03%.This method effectively complements the existing research on the trend prediction of internet public opinion,and provides support for the more accurate implementation of internet public opinion governance.
出处
《情报理论与实践》
CSSCI
北大核心
2020年第10期165-170,156,共7页
Information Studies:Theory & Application
基金
国家自然科学基金项目“异构信息网络下技术供需匹配模型与对接路径研究”的成果之一,项目编号:71974009。