Weighted Random Walk Sampling for Multi-Relational Recommendation

Vahedian, Fatemeh; Burke, Robin; Mobasher, Bamshad

doi:10.1145/3079628.3079685

Cited by 16 publications

(6 citation statements)

References 13 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Random walk strategy has been widely applied in recommendation. [53] performs random walk along the social network to search relevant users who have similar preference with the target user for better rating prediction; [54] exploits random walk to obtain diverse recommendation; [55] further extends [54] in heterogenous information network to generate valuable meta-paths; [54] performs random walk to complete implicit feedback matrix for mitigating data sparsity problem; Similarly, [25] employs random walk to find more positive instances. We remark that these works adopt static (uniform or pre-defined) transfer probability in their random walk strategy.…”

Section: Related Workmentioning

confidence: 99%

SamWalker++: recommendation with informative sampling strategy

Wang¹,

Chen²,

Zhou³

et al. 2020

Preprint

View full text Add to dashboard Cite

Recommendation from implicit feedback is a highly challenging task due to the lack of reliable negative feedback data. Existing methods address this challenge by treating all the un-observed data as negative (dislike) but downweight the confidence of these data. However, this treatment causes two problems: (1) Confidence weights of the unobserved data are usually assigned manually, which lack flexibility and may create empirical bias on evaluating user's preference. (2) To handle massive volume of the unobserved feedback data, most of the existing methods rely on stochastic inference and data sampling strategies. However, since a user is only aware of a very small fraction of items in a large dataset, it is difficult for existing samplers to select informative training instances in which the user really dislikes the item rather than does not know it. To address the above two problems, we propose two novel recommendation methods SamWalker and SamWalker++ that support both adaptive confidence assignment and efficient model learning. SamWalker models data confidence with a social network-aware function, which can adaptively specify different weights to different data according to users' social contexts. However, the social network information may not be available in many recommender systems, which hinders application of SamWalker. Thus, we further propose SamWalker++, which does not require any side information and models data confidence with a constructed pseudo-social network. In the pseudo-social network, similar users are connected with specific item nodes or community nodes. This way, the inference of one's data confidence can benefit from the knowledge from other similar users. We also develop fast random-walk-based sampling strategies for our SamWalker and SamWalker++ to adaptively draw informative training instances, which can speed up gradient estimation and reduce sampling variance. Extensive experiments on five real-world datasets demonstrate the superiority of the proposed SamWalker and SamWalker++.

show abstract

Section: Related Workmentioning

confidence: 99%

SamWalker++: recommendation with informative sampling strategy

Wang¹,

Chen²,

Zhou³

et al. 2020

Preprint

View full text Add to dashboard Cite

show abstract

“…However, the weights will greatly improve the information of the samples. For example, as shown in Figure 4, ifp (x i ) q(x i ) is equal to 1, then w(x i ) is equal to 1 according to equation (3), which means sample x i must subject to the target distribution. Therefore, the distribution of samples with higher weight is more approximate to the target distribution.…”

Section: E Importance Resamplingmentioning

confidence: 99%

“…However, how to further improve the accuracy of Web services QoS prediction is still a problem. Traditional researches [1]- [3] mainly focus on increasing the complexity of the predicting models for fixing the problem and simply assume that the probability distribution of QoS datasets is uniform. However, in the real world, QoS datasets tend to follow a complex distribution, that the sampled data (training data of the predicting models) based on such assumption is biased and leads inaccurate prediction.…”

Section: Introductionmentioning

confidence: 99%

Research on the Influence of Sampling Methods for the Accuracy of Web Services QoS Prediction

Jian

2019

IEEE Access

View full text Add to dashboard Cite

In recent years, as the number of Web services, increases dramatically, the personalized Web service recommendation has become a hot topic in both academia and industry. The quality-of-service (QoS) prediction plays a key role in Web service recommendation systems. However, how to further improve the accuracy of QoS prediction is still a problem. Traditional QoS predicting models do not consider the impact of sampling methods on the accuracy of QoS prediction. However, the outstanding sampling method can train the predicting model more effectively and obtain higher accuracy. Therefore, it is necessary to study sampling methods based on the QoS dataset in order to obtain sample distribution closer to the original distribution, so as to improve the accuracy of the predicting models. In this paper, we first discuss how to apply several existing sampling methods to QoS datasets and then analyze their advantages and disadvantages. Finally, a novel sampling method, enhanced importance resampling (EIRS), is proposed and applied. The experiments on the real-world datasets show that our method can not only sample efficiently and accurately but also can greatly improve the accuracy of Web service QoS prediction. INDEX TERMS Quality of service, Web services, sampling methods, enhanced importance resampling.

show abstract

“…Therefore, we should give higher weights to positions with the higher preferences. Based on reference [38], the userlocation edge weight is define as follows: u,p) ∀p ∈P u e r (u,p)…”

Section: Quantification Of Edge Weights Between Layers 1) User Prementioning

confidence: 99%

D2D Relay Communication Scheme Incorporating Multi-Dimensional Information in Multimedia Transmission

et al. 2019

View full text Add to dashboard Cite

The rapid development of mobile multimedia services integrated with online and offline social networks has made the mutual transmission of smart devices more frequent. Device to Device relay communication (D2D relay communication) can extend the transmission range of shared multimedia services, further reducing the burden on the base stations. However, when multimedia services are transmitted by D2D relay communication, the selfishness of the relay nodes affects the quality of the transmission. In order to solve this problem, the integration of social factors has become a common research topic. However, social factors generally come from multi-dimensional information. The more kinds of information, the more hidden information, and the measurement of social relations will become extremely difficult. Hence, this paper focuses on the measurement of social relationships based on multi-dimensional information and then combines physical domains to choose relay nodes. Firstly, the transmission performance of D2D relay communication is analyzed in the physical domain. Then, in order to better integrate multi-dimensional information, we construct a hypernet architecture based on spatiotemporal-user-location-category and quantify the edge weights. Three categories of weighted super-edge structures are defined to capture a variety of hidden social relationships, and the social relationship strength of any two users is measured. Finally, a relay selection algorithm that maximizes the social rate is proposed, which combines social relationships of multi-dimensional information and quality of physical transmission. The simulation results show that the Based on Maximum Social Rate for Relay Selection Algorithm (MSRRS) proposed in this paper has a significant improvement in transmission quality and transmission stability for multimedia application compared with the traditional relay selection algorithm. INDEX TERMS D2D, relay communication, social awareness, hypernet, multimedia transmission. How to meet people's growing demand for communication services with limited resources is an urgent problem to be solved [5]. D2D communication is a technology for communication of adjacent terminals [6], [7]. After the establishment of the D2D communication link between two adjacent terminals, data communication among devices will no longer need the core network's support or intermediate equipment [8], [9].

show abstract

Weighted Random Walk Sampling for Multi-Relational Recommendation

Cited by 16 publications

References 13 publications

SamWalker++: recommendation with informative sampling strategy

SamWalker++: recommendation with informative sampling strategy

Research on the Influence of Sampling Methods for the Accuracy of Web Services QoS Prediction

D2D Relay Communication Scheme Incorporating Multi-Dimensional Information in Multimedia Transmission

Contact Info

Product

Resources

About