Deep Collaborative Embedding for information cascade prediction

Zhao, Yuhui; Yang, Ning; Lin, Tao; Yu, Philip S.

doi:10.1016/j.knosys.2020.105502

Cited by 24 publications

(8 citation statements)

References 42 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…MAP has been used for evaluating cascading prediction [16,23,17,21,22], in this paper we also take MAP as the precision evaluation for prediction performance. The main idea of MAP is to set n cut-off according to the prediction order, calculate the prediction accuracy of each prediction top-k fragments, and finally compute the average accuracy.…”

Section: Prediction Metricmentioning

confidence: 99%

“…Topo-LSTM model [21] puts dynamic directed acyclic graphs into an LSTM-based model to generate topology-aware embeddings for nodes, which depends on network structure information. Deep collaborative embedding model [22] collaboratively embed the nodes with a deep architecture into a latent space, which can learn nodes' embeddings with the information of diffusion order. Yet, these embedding-based methods pay less attention on the interference when embedding nodes into latent spaces, which may lead that the embedding vectors of unrelated nodes are very close.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Independent Asymmetric Embedding for Information Diffusion Prediction on Social Networks

Xie,

Wang,

Jia

2021

Preprint

View full text Add to dashboard Cite

The prediction for information diffusion on social networks has great practical significance in marketing and public opinion control. Cascade prediction aims to predict the individuals who will potentially repost the message on the social network. One kind of methods either exploit demographical, structural, and temporal features for prediction, or explicitly rely on particular information diffusion models. The other kind of models are fully data-driven and do not require a global network structure. Thus massive diffusion prediction models based on network embedding are proposed. These models embed the users into the latent space using their cascade information, but are lack of consideration for the intervene among users when embedding. In this paper, we propose an independent asymmetric embedding method to learn social embedding for cascade prediction. Different from existing methods, our method embeds each individual into one latent influence space and multiple latent susceptibility spaces. Furthermore, our method captures the co-occurrence regulation of user combination in cascades to improve the calculating effectiveness. The results of extensive experiments conducted on real-world datasets verify both the predictive accuracy and costeffectiveness of our approach.

show abstract

Section: Prediction Metricmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Independent Asymmetric Embedding for Information Diffusion Prediction on Social Networks

Xie,

Wang,

Jia

2021

Preprint

View full text Add to dashboard Cite

show abstract

“…al. [11] investigated a prediction of both node infection and infection order without the knowledge about the underlying cascade model and the network. For this, they designed a novel model called Deep Collaborative Embedding (DCE) for information cascade prediction, which can capture not only the node structural property but also two kinds of node cascading characteristics such as cascading context and cascading affinity.…”

Section: Observationmentioning

confidence: 99%

“…It selects the node such that the Breadth-First Search (BFS) tree from the node has the minimum depth but the maximum number of leaf nodes. They also established some performance guarantees of SFT under the [9], LER [10] Ball & Tree [9], DCE [11], DeepDiffuse [12], GSLDA [13], Cascade-LSTM [14] Cascade Source Detection Not applicable…”

Section: Observationmentioning

confidence: 99%

Identification of Individual Infection Over Networks With Limit Observation: Random vs. Epidemic?

Choi

2021

IEEE Access

View full text Add to dashboard Cite

Popular information or dangerous viruses have recently been observed to spread rapidly through a highly connected network structure. For example, malicious viruses and rumors are spreading rapidly through online social network platforms over the Internet and diseases with high transmission power are rapidly spreading through human contact. Apart from these, some infections may occur individually regardless of the effect of the network such as computer failure. In the context of these infections coexisting, it is one of crucial problem to distinguish whether the infection with externally similar symptoms is caused by a cascade or by itself because if the infection is the cascade that is spreading through the network, it is necessary to stop this spread as soon as possible. In this paper, we study this classification problem to determine whether it is infected from a cascade or randomly when the infection snapshot is partially given. We propose two approaches for the problem (i) Neighbor-based approach as the use of local infection status information and (ii) Source-based approach as the global infection status information. The first one is that we just count the number of connected infection paths from L-hop distance to an infected node by using some criteria whereas, the second one is that we use the information of the location of cascade sources to infer the infection cause of each infected node by computing the infection probability from the sources. We perform various simulations to obtain the classification performance of the two proposed algorithms. As a result, the method of estimating a global cascade source shows better performance than that of the former one, which uses only the infection information of local neighboring nodes if the sampling rate of cascade infections is sufficient.

show abstract

“…[35] applies specifically designed graph neural network models to capture the change of node state as well as the network structure. [36] uses structural property and the order of cascaded nodes to predict the future sequence of cascades. [37] combines the representation of network and the time of retweet to predict the future spreading size.…”

Section: Related Workmentioning

confidence: 99%

CasSeqGCN: Combining Network Structure and Temporal Sequence to Predict Information Cascades

Wang¹,

Michalski²,

Ran³

et al. 2021

Preprint

View full text Add to dashboard Cite

One important task in the study of information cascade is to predict the future recipients of a message given its past spreading trajectory. While the network structure serves as the backbone of the spreading, an accurate prediction can hardly be made without the knowledge of the dynamics on the network. The temporal information in the spreading sequence captures many hidden features, but predictions based on sequence alone have their limitations. Recent efforts start to explore the possibility of combining both the network structure and the temporal feature for a more accurate prediction. Nevertheless, it is still a challenge to efficiently and optimally associate these two interdependent factors. Here, we propose a new end-to-end prediction method CasSeqGCN in which the structure and temporal feature are simultaneously taken into account. A cascade is divided into multiple snapshots which record the network topology and the state of nodes. The graph convolutional network (GCN) is used to learn the representation of a snapshot. The dynamic routing and the long short-term memory (LSTM) model are used to aggregate node representation and extract temporal information. CasSeqGCN predicts the future cascade size more accurately compared with other state-of-art baseline methods. The ablation study demonstrates that the improvement mainly comes from the design of the input and the GCN layer. Taken together, our method confirms the benefit of combining the structural and temporal features in cascade prediction, which not only brings new insights but can also serve as a useful baseline method for future studies.

show abstract

Deep Collaborative Embedding for information cascade prediction

Cited by 24 publications

References 42 publications

Independent Asymmetric Embedding for Information Diffusion Prediction on Social Networks

Independent Asymmetric Embedding for Information Diffusion Prediction on Social Networks

Identification of Individual Infection Over Networks With Limit Observation: Random vs. Epidemic?

CasSeqGCN: Combining Network Structure and Temporal Sequence to Predict Information Cascades

Contact Info

Product

Resources

About