Nonnegative matrix factorization algorithms for link prediction in temporal networks using graph communicability

Ma, Xiaoke; Sun, Penggang; Qin, Guimin

doi:10.1016/j.patcog.2017.06.025

Cited by 96 publications

(32 citation statements)

References 39 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Typically, CF recommends items to users by learning user/item similarities from existing ratings. A classical CF model is Matrix Factorization (MF) [8,9,10,11], which is a latent factor model [12]. It decomposes the rating matrix into two factor matrices, representing latent factors for users and items respectively.…”

Section: Introductionmentioning

confidence: 99%

Joint interaction with context operation for collaborative filtering

Bai

Liu

et al. 2019

Pattern Recognition

View full text Add to dashboard Cite

In recommender systems, the classical matrix factorization model for collaborative filtering only considers joint interactions between users and items. In contrast, context-aware recommender systems (CARS) use contexts to improve recommendation performance. Some early CARS models treat user, item and context equally, unable to capture contextual impact accurately. More recent models perform context operations on users and items separately, leading to "double-counting" of contextual information. This paper proposes a new model, Joint Interaction with Context Operation (JICO), to integrate the joint interaction model with the context operation model, via two layers. The joint interaction layer models interactions between users and items via an interaction tensor. The context operation layer captures contextual information via a contextual operating tensor. We evaluate JICO on four datasets and conduct novel studies, including varying contextual influence and time split recommendation. JICO consistently outperforms competing methods, while providing many useful insights to assist further analysis.

show abstract

Section: Introductionmentioning

confidence: 99%

Joint interaction with context operation for collaborative filtering

Bai

Liu

et al. 2019

Pattern Recognition

View full text Add to dashboard Cite

show abstract

“…On the other hand, the dynamics prediction problem of network systems can be generally modeled as the temporal link prediction task, and a brief overview about the task can be found in [12] and [13].…”

Section: Related Workmentioning

confidence: 99%

“…To quantitatively evaluate our temporal link prediction model, we follow previous work [6], [12] to use the Mean Square Error (MSE) scores for comparison, which is defined as:…”

Section: B Evaluation Metricsmentioning

confidence: 99%

“…We evaluate the effectiveness of our model by comparing it with six baseline methods on the four datasets, including ED [12], SVD [12], NMF [8], GrNMF [6], AM-NMF [17] and LSTM [10]. Among the baselines, ED, SVD and NMF are conventional approaches based on the collapsed network [14], [15], while GrNMF and AM-NMF are state-of-the-art matrix factorization-based methods without collapsing the dynamic network.…”

Section: Performance Evaluationmentioning

confidence: 99%

“…The parameter settings of the four datasets are shown in Table IV, where λ is the parameter controls the effect of the L 2 -regularization term in (12); ε is the threshold used to refine the prediction result in (17); α 0 is the learning rate of the pretraining process; α D and α G are the learning rate for training D and G; c is the clipping bound for D's parameters θ D . For each dataset, we use the first (l + 2) graph snapshots as the original training data to conduct the initial training process.…”

Section: Performance Evaluationmentioning

confidence: 99%

See 2 more Smart Citations

GCN-GAN: A Non-linear Temporal Link Prediction Model for Weighted Dynamic Networks

Lei

Qin

Bai

et al. 2019

IEEE INFOCOM 2019 - IEEE Conference on Computer Communications

169

View full text Add to dashboard Cite

In this paper, we generally formulate the dynamics prediction problem of various network systems (e.g., the prediction of mobility, traffic and topology) as the temporal link prediction task. Different from conventional techniques of temporal link prediction that ignore the potential non-linear characteristics and the informative link weights in the dynamic network, we introduce a novel non-linear model GCN-GAN to tackle the challenging temporal link prediction task of weighted dynamic networks. The proposed model leverages the benefits of the graph convolutional network (GCN), long short-term memory (LSTM) as well as the generative adversarial network (GAN). Thus, the dynamics, topology structure and evolutionary patterns of weighted dynamic networks can be fully exploited to improve the temporal link prediction performance. Concretely, we first utilize GCN to explore the local topological characteristics of each single snapshot and then employ LSTM to characterize the evolving features of the dynamic networks. Moreover, GAN is used to enhance the ability of the model to generate the next weighted network snapshot, which can effectively tackle the sparsity and the wide-value-range problem of edge weights in real-life dynamic networks. To verify the model's effectiveness, we conduct extensive experiments on four datasets of different network systems and application scenarios. The experimental results demonstrate that our model achieves impressive results compared to the state-of-the-art competitors.

show abstract