基于并行决策树的微博互动数预测

Interaction number prediction of micro-blog based on parallel decision tree

社交网络的快速发展,微博成为主要的社交媒体平台,针对如何预测微博文本的未来互动数,对微博进行有效的分发控制的问题,提出一种基于并行决策树的微博互动数所属级数预测的方法。首先,对用户以往发表的微博进行用户特征和微博文本特征的处理;然后,使用并行决策树分类算法对训练数据进行分类模型的构建;最后使用得到的分类模型对新微博文本的互动数所属级数进行分类预测。通过对比算法的实验,验证了所提方法具有较高的分类精度和较好的可扩展性,能够对微博所属级数进行有效的分类预测。

To predict the future interaction number of micro-blog texts to implement effective distribution control of micro-blogs, a method of forecasting the series number of micro-blog interaction numbers based on parallel decision tree was proposed. Firstly, the user characteristics and micro-blog text features of the user＇ s previous miero-blog were processed. Then, a classification model of the training data was constructed via a parallel decision tree classification algorithm. Finally, the series number of the interaction number of new micro-blog texts was classified via the classification model. The experimental results show that the proposed method has high classification accuracy and good scalability and can effectively forecast micro-blog series.