A comparative study of heterogeneous item recommendations in social systems

Bellogín, Alejandro; Cantador, Iván; Castells, Pablo

doi:10.1016/j.ins.2012.09.039

Cited by 67 publications

(37 citation statements)

References 55 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…A similar approach for aggregating data from multiple OSNs was presented in [10], where the dataset was enriched by semantics from external resources, like DBpedia. In [3], data from multiple OSNs was analysed in order to find what information could improve the diversity of recommendations. In [12], the authors investigated the feasibility and effectiveness of using cross-domain data for generating Facebook recommendations.…”

Section: Introductionmentioning

confidence: 99%

Cross social networks interests predictions based ongraph features

Tiroshi

Berkovsky

Kâafar

et al. 2013

Proceedings of the 7th ACM Conference on Recommender Systems

View full text Add to dashboard Cite

The tremendous popularity of Online Social Networks (OSN) has led to situations, where users have their profiles spread across multiple networks. These partial profiles reflect different user characteristics, depending mainly on the nature of the network, e.g., Facebook's social vs. LinkedIn's professional focus. Combining data gathered by multiple networks may benefit individual users, and the community as a whole, as this could facilitate the provision of more accurate services and recommendations. This paper reports on an exploratory study of the process of making such recommendations using a unique multi-network dataset containing user interests across multiple domains, e.g., music, books, and movies. We represent the data using a graph model and generate recommendations using a set of features extracted from and populated by the model. We assess the contribution of various network-and domain-related features to the accuracy of the recommendations and motivate future work into automated feature selection.

show abstract

Section: Introductionmentioning

confidence: 99%

Cross social networks interests predictions based ongraph features

Tiroshi

Berkovsky

Kâafar

et al. 2013

Proceedings of the 7th ACM Conference on Recommender Systems

View full text Add to dashboard Cite

show abstract

“…It inevitably suffers from sparse and imbalance of rating data. Alejandro Bellogłn et al [17] get the conclusions that when explicit social networks are available, incorporating characteristics of social graphs into the computation of user neighbors in memory-based CF significantly improves recommendation in terms of ranking quality and social tagging can easily be exploited to provide precise item recommendation ranking lists. According to this, we fuse social relations and user-generated tags into CF to improve the accuracy of recommendations.…”

Section: Proposed Methodsmentioning

confidence: 99%

“…However, depending on the nature of a system, it may be difficult to find users with enough information of each type [17]. So in our paper, we choose the Last.fm dataset [19] which could meets our needs after simple preprocessing.…”

Section: Datasetsmentioning

confidence: 99%

See 1 more Smart Citation

Social Network and Tag Sources Based Augmenting Collaborative Recommender System

Zhou

Tang

et al. 2015

IEICE Trans. Inf. & Syst.

251

View full text Add to dashboard Cite

SUMMARYRecommender systems, which provide users with recommendations of content suited to their needs, have received great attention in today's online business world. However, most recommendation approaches exploit only a single source of input data and suffer from the data sparsity problem and the cold start problem. To improve recommendation accuracy in this situation, additional sources of information, such as friend relationship and user-generated tags, should be incorporated in recommendation systems. In this paper, we revise the user-based collaborative filtering (CF) technique, and propose two recommendation approaches fusing usergenerated tags and social relations in a novel way. In order to evaluate the performance of our approaches, we compare experimental results with two baseline methods: user-based CF and user-based CF with weighted friendship similarity using the real datasets (Last.fm and Movielens). Our experimental results show that our methods get higher accuracy. We also verify our methods in cold-start settings, and our methods achieve more precise recommendations than the compared approaches.

show abstract

“…Although some researchers have designed recommendation models [22] to give consumers advice on products and/or service selection based on some social information, most of them considered only one single trust value in their recommendation system. Moreover, there is still a considerable gap between their models and the real-social networks and it is still a long way to go to represent the trust level of a service provider more comprehensively.…”

Section: B Service Discovery/selection In Osnsmentioning

confidence: 99%

Trust Evaluation via Large-Scale Complex Service-Oriented Online Social Networks

Liu

Jia

2015

IEEE Trans. Syst. Man Cybern, Syst.

View full text Add to dashboard Cite

Service-oriented online social networks (SOSNs) are emerging ubiquitous platforms for numerous services where service consumers require the selection of trustworthy service providers who are unknown to them before invoking services with the aid of other intermediate participants. Under this circumstance, evaluation of the trust level of the service provider along the social trust paths from the service consumer to the service provider is required. To this end, selection of the optimal social trust path (OSTP) that can yield the most trustworthy evaluation result is a prerequisite. While existing single-trust-value methods can provide good but simple information to service consumers, more trust information, such as social intimacy degree between participants and role impact factor of intermediate participants, should be considered to represent the trust level of a service provider more comprehensively. When more trust information is considered, OSTP selection will become an NP-complete problem. In this paper, we propose path integral Monte Carlo quantum annealing (PIMCQA)-based OSTP (PIMCQA_OSTP) selection algorithm for complex SOSNs. PIMCQA_OSTP serves as the very first quantum inspired OSTP selection algorithm in complex SOSNs. Due to that quantum mechanics work with wave functions that can sample different regions of phase space equally well, and quantum systems can tunnel through classically impenetrable potential barriers between energy valleys, PIMCQA_OSTP shows its outstanding search ability and outperforms existing methods. Results of experiments on a real dataset of online social networks verify that PIMCQA_OSTP is a promising tool and is especially fit for complex SOSNs.Index Terms-Complex service-oriented online social networks, optimal social trust path selection (OSTP), path integral quantum annealing, quantum tunnel, trust level.

show abstract

A comparative study of heterogeneous item recommendations in social systems

Cited by 67 publications

References 55 publications

Cross social networks interests predictions based ongraph features

Cross social networks interests predictions based ongraph features

Social Network and Tag Sources Based Augmenting Collaborative Recommender System

Trust Evaluation via Large-Scale Complex Service-Oriented Online Social Networks

Contact Info

Product

Resources

About