社交网络中微博用户行为的分析与预测

Analysis and Prediction of Microblog User Behavior on Social Networks

从真实的新浪微博数据中分析用户的转发行为,提取了用户特征、微博特征、交互特征和结构特征等4个方面的影响转发行为的因素。通过实证分析各个特征对转发行为的具体影响,并利用机器学习中的不同预测算法对用户是否会对给定主题的微博产生转发行为进行预测。实验表明,用我们选取的因素,结合逻辑回归模型对于用户转发行为的预测更加准确。

Microblog is a platform on which the public can put forward their opinions and communicate with each other.Retweet behavior is the key mechanism for information diffusion on microblogging networks.On the basis of real data from Sina Weibo,this paper predicts users＇ retweet behavior using four features which affect user＇s behavior,including user attributes,microblog text features,interactive attributes and local structures.Furthermore,we utilize different supervised classifiers to predict retweet behavior.The experiments and evaluation show the effectiveness of our feature choices.The results show that the selected,features combined with the Logistic Regression model,can predict users＇ sretweet behavior more accurately.