基于双分支结构的融合多特征微博传播行为预测算法

随着如今微博平台的高速发展，微博转发行为预测已经成为舆情分析领域中一个热门的研究主题。针对该任务，提出一种添加多层间接粉丝用户权威度信息，基于双分支网络结构模型的微博转发行为预测算法。该方法通过对原始微博进行分析，运用LDA算法提取内容特征、构建用户关系网络提取间接关注用户权威度特征等多元特征，构建基于双分支结构神经网络模型预测微博传播行为。实验结果表明预测模型相比于其他算法在RMSE,MAE评估指标上都有较好的提高，验证了算法的有效性。

Microblog Propagation Behavior Prediction Algorithm with Multiple Features Based on Double Branch Structure

With the rapid development of microblog platforms nowadays,Weibo forwarding behavior prediction has become a hot research topic in the field of public opinion analysis.Aiming at this task, this paper proposes a microblog forwarding behavior prediction algorithm based on the dual-branch network structure model. The method analyzes the original microblog and uses the LDA algorithm to extract the content features, constructs the user relationship network to extract the authority feature about indirect relationship user and other multi-features, then constructs the dual-branch structure neural network model to predict the micro-blog propagation behavior. The experimental results show that the prediction model has better improvement on the RMSE and MAE value than other algorithms, and verify the effectiveness of the algorithm.