论文部分内容阅读
以短文本内容发布为主要特点的微博,已经成为重要的信息传播媒介,预测微博流行度对舆情监测、企业营销、热点推送等都具有重要意义.当前对微博流行度预测的研究主要侧重于对所有用户的微博数据进行统一建模预测,鲜有研究考虑不同影响力用户之间的差异.而微博数据的分析显示标签、提及和微博长度等对微博流行度的影响会随发布者的影响力变化显示出明显差异,在流行度预测中充分考虑这些差异,有助于取得更好的预测结果.为此,在流行度预测中引入多任务学习(Multi-Task Learning,简称MTL),并结合SVM构建SVM+MTL模型,此模型通过同时考虑所有用户的共同特性和不同用户的具体特性来提高预测性能.此外,除了预测常用的用户属性和微博发布行为等特征外,还引入微博内容相似性这一新特征,该特征能明显提高预测准确率.基于微博数据的实验表明,SVM+MTL模型可以有效提高微博流行度预测性能.
Weibo, which has the short text content as its main feature, has become an important information media, and it is of great significance to predict the popularity of Weibo to public opinion monitoring, corporate marketing, hot push, etc. The current research on the prediction of the popularity of Weibo Focusing on the unified modeling and forecasting of all the users’ microblogging data and few studies considering the differences among users with different influences.An analysis of the microblogging data shows that the tagging, mentions, The impact will vary significantly with the publisher’s influence, and taking these differences into account in predicting the prevalence will help to achieve better predictions. To this end, Multi-Task Learning (MTL for short), combined with SVM to build SVM + MTL model, this model by considering all the common characteristics of all users and the specific characteristics of different users to improve the predictive performance.In addition, in addition to predicting common user attributes and microblogging release behavior Features, but also the introduction of the new features of the similarity of the content of the microblogging, which can significantly improve the prediction accuracy.Experiments based on Weibo data show that the SVM + MTL model To improve the popularity of microblogging to predict performance.