基于 SOM 神经网络的旅游突发事件网络舆情的传播态势

摘要

First, We selected 30 hot tourism emergency events occurred from 2010 to 2014 as a the research ob-ject, write program to crawl microblog data on the microblog API platform and used SOM neural network toolbox in Matlab 7.1 to cluster microblog data.Second, we chose a result type and used the exponential function method for forecasting the spread trend.The result of the paper show that, microblog spread trend of tourism emergency events is mainly divided into six types, there are differences characteristics between different types; the predictive effect of microblog spread trend of sixth clustering result is good.%微博平台作为网络舆情传播渠道的典型代表，拥有强大的聚合民意、组织和分配资源能力，同时还有助长旅游突发事件的危机态势爆发的劣势。选取2010－2014年内发生的30个热点旅游突发事件作为研究对象，采用新浪微博API平台抓取微博数据，调用Matlab7畅1的SOM神经网络工具箱进行聚类，并采用指数函数方法进行拟合预测。研究结果表明，旅游突发事件的网络舆情传播态势主要分为六种类型，不同类型存在传播特征差异；对第六聚类旅游突发事件的微博传播态势的预测效果良好，与实际传播态势吻合。