Quantitative predicting propagation breadth and depth of microblog users’ forwarding behavior

Abstract

In the microblog network, users’ forwarding behavior is widespread and the propagation range is difficult to predict quantitatively. To solve this problem, machine learning algorithms are used to quantitatively predict propagation breadth and depth of microblog users’ forwarding behavior. The dataset is preprocessed, and the extracted features are divided into three types: user features, microblog features and social features. Then the dataset is analyzed in detail; machine learning algorithms are used to predict the propagation breadth and depth of users’ forwarding behavior; and the influence of the three types of features on prediction precision is studied. The experimental results show that the prediction precision of the improved random forest algorithm has less fluctuations, and it is not sensitive to the changes of various features. The improved random forest algorithm has higher precision and better generalization ability than the other algorithms, which shows that the prediction results have high reference value. Social features have the greatest influence on the prediction precision for each prediction algorithm. User features have the similar influence as microblog features on the prediction precision.