Information Cascades Prediction With Graph Attention

Abstract

The cascades prediction aims to predict the possible information diffusion path in the future based on cascades of the social network. Recently, the existing researches based on deep learning have achieved remarkable results, which indicates the great potential to support cascade prediction task. However, most prior arts only considered either cascade features or user relationship network to predict cascade, which leads to the performance limitation because of the lack of unified modeling for the potential relationship between them. To that end, in this paper, we propose a recurrent neural network model with graph attention mechanism, which constructs a seq2seq framework to learn the spatial-temporal cascade features. Specifically, for user spatial feature, we learn potential relationship among users based on social network through graph attention network. Then, for temporal feature, a recurrent neural network is built to learn their structural context in several different time intervals based on timestamp with a time-decay attention. Finally, we predict the next user with the latest cascade representation which obtained by above method. Experimental results on two real-world datasets show that our model achieves better performance than the baselines on the both evaluation metrics of HITS and mean average precision.

An Online Topic Diffusion Prediction Approach on Heterogeneous Social Networks

“…IEPPA was used to deduce the information propagation in the latent embedding space according to propagation prediction model depending on information. RNEPPPA was applied for precisely predicting the information diffusion popularity only ≈74% PA: performance of A, PP: performance of P: unimodal, U: unimodal, M: multimodal, E: the total number of topics prediction algorithm [14] (abbreviated as IEPPA), repost network embedding-based propagation popularity prediction algorithm [22] (abbreviated as RNEPPPA), recurrent neural network model [23] (abbreviated as RNNM) and conditional probability model of information diffusion [29] (abbreviated as IDCPM). IEPPA was used to deduce the information propagation in the latent embedding space according to propagation prediction model depending on information.…”

“…RNe2Vec (repost network to vector), which is a repost network embedding-based diffusion popularity prediction algorithm, predicted the information diffusion popularity only based on early repost information, given that the underlying user relation network precisely [22]. Then a recurrent neural network model with graph attention mechanism, which constructs a seq2seq framework to learn the spatial-temporal cascade features, was built to learn their structural context in several different time intervals based on timestamp with a time-decay attention and predict the next user with the latest cascade representation [23]. A formal model was developed to characterize the dynamic process of knowledge diffusion in the autonomous learning under multiple networks to expand the scope of knowledge through hybrid online learning for allocating educational resources [24].…”

an Online Topic Diffusion Prediction Approach on Heterogeneous Social Networks

“…To validate our methods' effectiveness and efficiency, we aim to compare the application scope of the topic diffusion trend prediction, the prediction precision including the determined coefficient and precision among the unimodal topic diffusion trend prediction method (abbreviated as LUTDTPM), short-term multimodal topic diffusion prediction method (abbreviated as SMTDPM), information-dependent embedding-based propagation prediction algorithm [14] (abbreviated as IEPPA), repost network embedding-based propagation popularity prediction algorithm [22] (abbreviated as RNEPPPA), recurrent neural network model [23] (abbreviated as RNNM) and conditional probability model of information diffusion [29] (abbreviated as IDCPM). IEPPA was used to deduce the information propagation in the latent embedding space according to propagation prediction model depending on information.…”

“…RNe2Vec (repost network to vector) [22] , which is a repost network embedding-based diffusion popularity prediction algorithm, predicted the information diffusion popularity only based on early repost information, given that the underlying user relation network precisely. Then a recurrent neural network model [23] with graph attention mechanism, which constructs a seq2seq framework to learn the spatial-temporal cascade features, was built to learn their structural context in several different time intervals based on timestamp with a time-decay attention and predict the next user with the latest cascade representation. A formal model [24] was developed to characterize the dynamic process of knowledge diffusion in the autonomous learning under multiple networks to expand the scope of knowledge through hybrid online learning for allocating educational resources.…”