基于动量模型的食品安全事件发现方法

Discovery method of food safety incidents based on momentum model

食品安全是广受民众关注的热点话题,而微博已经成为食品安全事件曝光的主要媒体平台.以微博语料作为数据源,同时使用微博内容和用户的社交网络行为特征,提出了基于动量模型的食品安全事件发现方法.该方法以事件发现作为描述食品安全事件的基本模型,以检测出微博信息流中与食品安全相关的候选特征词,然后采用动量模型实现候选特征词的动量建模和重复特征词的有效过滤.最后,通过K-means聚类将有效的特征词进行归类合并,以实现食品安全事件的发现.试验结果表明:该方法能够有效发现在微博中传播的食品安全事件,并能过滤掉微博中无关的话题.

As a hot topic,food safety has attracted a lot of attention from the public,and microblog has become the main media platform to expose food safety incidents. Microblog corpus was used as data source with microblog content and user social network behavior characteristics,and the food safety incident discovery method was proposed based on the momentum model. To describe the food safety incident from microblog information flow,the event discovery model was used to detect the candidate feature words related to food safety. The momentum model was established to realize the momentum modeling of candidate feature words and filter the duplicate feature words effectively. The effective feature words were classified and merged by K-means clustering,and the goal of discovering food safety incidents was achieved. The experimental results show that the proposed method can effectively discover the food safety incidents spreading in microblog and filter out irrelevant topics in microblog.