针对多数研究仅将社会化媒体作为数据来源的现状,深入分析社会化媒体特点,重点将节点属性分为静态和动态进行研究,提出基于预测目标的节点影响力的概念.在此基础上提出了一种基于节点属性进行信息预测的属性、节点数、倾向(ANV)模型.实验采用后向传播(BP)神经网络预测方法,通过新浪微博数据预测电影票房.仿真表明,带有节点属性的方法比没有节点属性的方法拟合和预测更为准确. In view of the fact that most of the researches only use social media as data source, this paper analyzes the characteristics of social media in detail, and focuses on the study of node attributes in static and dynamic terms and puts forward the concept of node influence based on forecasting goals. An Attribute, Number of Nodes, Propensity (ANV) Model for Information Prediction Based on Node Attributes The BP neural network prediction method is used in the experiment to predict the movie box office by using Sina Weibo data.The simulation results show that with attribute of node Methods are more accurate than method fitting and prediction without node properties.