基于微信用户行为的分享预测模型研究  

Prediction Model of Sharing Behavior Based on WeChat User

作　　者：黄炜[1,2] 余辉[1] 李岳峰[1] 

机构地区：[1]湖北工业大学经济与管理学院,湖北武汉430064 [2]武汉理工大学管理学院,湖北武汉430072

出　　处：《情报理论与实践》2016年第11期89-94,共6页Information Studies:Theory & Application

基　　金：国家自然科学基金项目"微博环境下实时主动感知网络舆情事件的多核方法研究"(项目编号:71303075);国家自然科学基金项目"大数据环境下基于特征本体学习的无监督文本分类方法研究"(项目编号:71571064)的成果之一

摘　　要：[目的/意义]构建微信用户分享行为的预测模型,探究微信分享的传播规律。[方法/过程]从用户行为动机、基本特征、群体传播特征3个维度分析并对用户的分享行为进行预测建模,并以微信热议"穹顶之下"为例,进行分享预测实例分析。[结果/结论]微信用户行为与现实人际关系有着较大的相关性,其行为受通信录好友、身边熟人、公众平台的影响,实例检测结果符合实际用户分享行为分布。其分享行为可以基于用户行为特征集进行预测,并能够达到一定的准确率。[局限]调查数据源和范围存在一定的局限性,给预测工作带来了一定的误差,未来可以加入更多维度和扩大调查范围来得到更准确的结果。[ Purpose/significance] This paper aims to construct the prediction model of WeChat users＇ sharing behaviors and explore the propagation pattern of WeChat sharing. [ Method/process ] The paper analyzes users＇ sharing behaviors through their behavior motivation, basic features and group spread features and constructs the prediction model of users＇ sharing behavior. The model has been tested by the instance of ＂Under the Dome＂ which is a hot topic on WeChat. [ Result/conclusion ] There is a strong connection between WeChat users＇ behavior and real-world interpersonal relationship. Users＇ behavior is influenced by address list, acquaintances and the public platform. The instance test results conform to the actual users＇ sharing behavior distribution. WeChat users＇ sharing behavior can be predicted based on users＇ behavior feature sets, which could achieve a certain accuracy rate. [ Limitations] Survey data source and scope are insufficient for the prediction. The study needs more dimensions and expanding scope of investigation to get more accurate results in the future.

关 键 词：微信 用户行为 信息分享 行为预测 

分 类 号：G206[文化科学—传播学] G252