With the surge of social media, micro-blog has become a popular information share tool, in which retweeting is a basic way to share and spread information. It is important to predict the retweet number for influence measure and precision market. Contemporary methods usually consider it as a classification or regression problem directly, which can be regarded as one-phase models. However, they cannot accurately predict the number of retweet. In this paper, we propose a two-phase model to predict how many times a tweet can be retweeted in Sina Weibo. That is, the model first classifies tweets into several categories, and then does regression on each category. Extensive experiments on real Sina Weibo dataset show that our model is a general framework to achieve better performances than traditional one-phase prediction model without complex feature extraction.

A Two-Phase Model for Retweet Number Prediction

Micro-blog has become the most popular information sharing tool in our daily life. The retweet behavior is a main method of information propagation in micro-blog. So there tweet number prediction not only is an interesting research topic, but also has much practical significance. However, most of current researches only regard this problem as a classification or regression problem, and they did not consider the retweet propagation process. In this paper, considering the retweet propagation process, we propose a retweet number prediction model BCI. In our model, we think retweet messages are from two parts, direct followers and indirect followers. Moreover, the retweet number of followers is decided by their retweet intention and influence. We use behavior and content information to estimate retweet intention for a direct follower and use the influence to estimate the indirect followers' retweet number. Experimental results on Sina Weibo dataset show that our retweet number prediction model has much better performance than other well-established methods.

A Retweet Number Prediction Model Based on Followers' Retweet Intention and Influence

Research on the diffusion effect of micro-blog plays an important role in improving marketing efficiency, strengthening monitoring public opinion and accurately discovering hotspot etc. To solve the problems not taking users' differences into consideration in the previous research, this paper proposes an algorithm to predict scale and depth of retweet massages based on analysis of retweet behavior. With the combination of LR algorithm and nine related features extracted from micro-blog users themselves, their relationships and micro-blog contents, we proposes a prediction model of retweet behavior. Based on this model, we proposes an algorithm to predict the diffusion effect, which considers the character of information spreading along users and does statistical analysis of adjacent users iteratively. Experimental results on Sina micro-blog dataset show that the algorithm has a prediction accuracy of 87.1% and 81.6% in scale and depth respectively, which indicates the model works well.

Research of micro-blog diffusion effect based on analysis of retweet behavior

Abstract This study aims at constructing a microblog influence prediction model and revealing how the user, time, and content features of microblog entries about public health emergencies affect the influence of microblog entries. Microblog entries about the Ebola outbreak are selected as data sets. The BM25 latent Dirichlet allocation model (LDA-BM25) is used to extract topics from the microblog entries. A microblog influence prediction model is proposed by using the random forest method. Results reveal that the proposed model can predict the influence of microblog entries about public health emergencies with a precision rate reaching 88.8%. The individual features that play a role in the influence of microblog entries, as well as their influence tendencies are also analyzed. The proposed microblog influence prediction model consists of user, time, and content features. It makes up the deficiency that content features are often ignored by other microblog influence prediction models. The roles of the three features in the influence of microblog entries are also discussed.

An Influence Prediction Model for Microblog Entries on Public Health Emergencies

In microblog networks, when a user posts a microblog, other users may forward the post, and then the forwarding process will bring about the rapid dissemination and diffusion of information. In this paper, we propose a comprehensive and novel approach to predict user forwarding behavior. Firstly, we build the feature sets that affect the microblog forwarding, such as interest topic,geographic location, user aggregation coefficient,neighborhood overlap and so on. These features are classified into four categories: user characteristics, microblog features, network structure features, and interactive behavior characteristics. Secondly, we establish a feature selection model based on Filtering and Wrapping for predicting the forwarding behavior of users. The model includes three aspects: (1)ANOVA(Analysis of variance): The value of each feature is analyzed by variance analysis. If the feature variance is small, the feature provides less information. (2)$\chi ^{2}$ test and point-two-column correlation analysis: They filter discrete and continuous features, respectively. (3)Wrapper analysis: In order to solve the strong correlations between the features, we use LVW(Las vagas wrapper) algorithm to analyze the above feature sets, and then obtain the optimal feature combination. Finally, we propose the forwarding prediction model based on AdaBoost(Adaptive boosting) algorithm. Experimental results demonstrate that the model has the highest precision and F1 score than Naive Bayes, Logistic Regression, Random Forest and SVM(Support vector machine), and the F1 score reached 0.885. Among different topics, our proposed AdaBoost prediction model has good recall and F1 scores for different topics. In addition, by using different feature sets for comparison experiments, it is found that the optimal features selected in this paper are very effective.

/pdf/forwarding-behavior-prediction-based-on-microblog-user-3lsmsqzqud.pdf

Forwarding Behavior Prediction Based on Microblog User Features

Based on the analysis of the factors that affect retweet behavior,this paper proposed a predict SVM algorithm with five features:user influence,user activity,interest similarity,the importance of micro-blog content and users closeness.Furthermore,it proposed the predict algorithm of retweet scale on the basis of SVM,also,gave a method to evaluate the predict accuracy.The experiment with Sina micro-blog data shows a good result that the predict accuracy is up to 86.63%.

Liu Li-xiong

Papers

Predict algorithm of micro-blog retweet scale based on SVM