摘要
作为重要的社会化媒体,微博凭借其社会性、媒体性、广泛参与性和快速传播4大特性已成为网络舆论的主要载体。针对品牌丑闻事件在微博上的传播进行研究,选取了29个近年来在微博上有明显传播特征的品牌丑闻事件作样本,以两小时为间隔搜集事件爆出后的一周时间内的微博博文数量,记录丑闻事件传播动态变化,并运用神经网络SOM模型方法对博文的数量变化进行聚类,得到微博上品牌丑闻事件传播的五种类型:对数型、缓坡型、突变型、长坡型及堤坝型,并在Matlab软件中用指数拟合的方法对数据做预测分析。企业了解丑闻事件在微博上的不同传播类型及各自的特点,对于其在快速预测和有效应对丑闻事件以及品牌危机时,具有很好的参考和借鉴价值。
As an important social media, microblogging has social, media and participatory extensive features, thereby enabling the rapid spread of public opinion, and it has increasingly become the main carrier for the network public opinion. We selected 29 brand scandal e- vents having significant propagation characteristics on the microblogging platform in recent years as a sample, collected the number of mi- croblogs within a week after the event broke, and thus recorded the dynamic changes of those events. Using neural network SOM model to find the change in the number of microblogs through clustering, brand scandal spread on the microblogging was categorized into five classi- fications. Exponential fitting method was used to get prediction of the data. Enterprises should understand the evolving history of scandals in the form of microblogs, their different types of transmission, and their respective characteristics. It may help them to respond more pro- actively to the scandals and brand crises.
出处
《情报杂志》
CSSCI
北大核心
2013年第10期23-28,12,共7页
Journal of Intelligence
基金
四川省2013年哲学社会科学重点项目"大型工程建设与环境污染事件的社会风险防范及应对机制研究"(编号:SC13A002)
关键词
品牌丑闻
微博
舆情监测
SOM神经网络
聚类
指数拟合
brand scandal microblogging moitoring public opinion SOM neural network clustering exponential fitting
