A Community-Based Recommendation System to Reveal Unexpected Interests

Kamahara, Junzo; Asakawa, T.; Shimojo, Shinsuke; Miyahara, Hideo

doi:10.1109/mmmc.2005.5

Cited by 38 publications

(17 citation statements)

References 15 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…no rating data). Some researchers [7,12] have proposed to incorporate community detection techniques in the context of recommender systems as a solution to overcome the above challenges and to make easier a future identification of the most related members or items which will be exploited to estimate preference prediction. In fact, by restricting the recommendation task on the communities into which the target item or user belongs instead to the whole network, not only the scalability of recommender systems may be improved, but also the data sparsity may be reduced.…”

Section: Recommender Systemsmentioning

confidence: 99%

A Dynamic Community-Based Personalization for e-Government Services

Abdrabbah

Ayachi

Amor

2016

Proceedings of the 9th International Conference on Theory and Practice of Electronic Governance

View full text Add to dashboard Cite

The government portals and websites contain a wide range of information and services to help citizens and businesses comply in ease with the government requirements. One of the most important challenges for e-Government services is matching particular needs and interests of citizens and businesses to achieve efficient eGovernment services quality. Recommender systems have been proposed to deliver users with the most interesting information service thereby addressing the information overload problem. In this paper, we focus on the implementation of a recommender system in an e-Government context to provide personalized G2C eservices. The proposed approach combines the expert pre-defined categorization of the e-government services and the citizens rating history to identify the relevant services for the active citizen. The efficiency of the proposed e-Government services personalization method is studied via a comparative study with the benchmark item-based and a state-of-the-art recommender system for Government-to-Business e-services. Experimental results show a considerable improvement for the proposed recommendation approach in term of performance and accuracy.

show abstract

Section: Recommender Systemsmentioning

confidence: 99%

A Dynamic Community-Based Personalization for e-Government Services

Abdrabbah

Ayachi

Amor

2016

Proceedings of the 9th International Conference on Theory and Practice of Electronic Governance

View full text Add to dashboard Cite

show abstract

“…The wandering process starts with a Google search of randomly chosen words from the profile in order to select pages that have more cross-domain integration, which allow to spark new interests. A similar approach is described in (Kamahara, Asakawa, Shimojo, & Miyahara, 2005), in which a method for locating unexpected items from clusters similar to a user cluster (the bubble) has been explored. The AURALIST framework for music recommendation is also based on the same strategy (Zhang et al, 2012).…”

Section: Programming For Serendipitymentioning

confidence: 99%

An investigation on the serendipity problem in recommender systems

Gemmis

Lops

Semeraro

et al. 2015

Information Processing & Management

105

View full text Add to dashboard Cite

“…Nor history of interest by of the user action was considered. In the research [5], recommendation function based on the user interest is introduced. The result of recommendation from the user action and the relations to the other user is presented to the user.…”

Section: Related Workmentioning

confidence: 99%

“…On the other hand, as progress of computing power and multimedia technology, application systems can use high quality 3D graphics. As a result, a large scale collaborative virtual environments (CVE) to support co-operation by many users over high-speed network have been proposed so far [1] [5]. However, it is required to guarantee scalability and usability of CVE system and usability to enhance CVE communication function as the number of participants increases from more several hundreds to thousand people using high quality video and audio communication facility.…”

Section: Introductionmentioning

confidence: 99%

Sclable Collaborative Virtual Environment Considering User's Interests Based on P2P Overlay Network

Nakai

Shibata

2007

First International Conference on Complex, Intelligent and Software Intensive Systems (CISIS'07)

View full text Add to dashboard Cite

Recently, many applications of VR called Collaborative Virtual Environment (CVE) have been proposed and their cooperation methods by users and scalability issues to support a number of users have been investigated. However, by using conventional client/server model, the usability of co-operative works by users degreases and scalability in CVE cannot be attained as the number of the user increases. Also, in conventional CVE system, the presentation and interaction of 3D objects in CVE to users were simple, limited and did not consider the user's background such as user's interest and generation. Eventually the user satisfaction for CVE application was not necessarily enough. In order to overcome those difficulties, we introduce a new method which uses P2P overlay network to realize scalable CVE and recommendation system which uses user's interest to present and offer the proper objects in CVE. Using our suggested method, the 3D objects are distributed on the P2P overlay networks and their computing loads can be reduced, eventually the scalability of the CVE can be improved. Furthermore, by reflecting user interest information to the application system, user can access and communicate to 3D 2objects and contents in which he or she is more interested with higher priority. Thus, the total usability of the CVE system can be improved. In this paper, we propose a new CVE system using P2P overlay network and user recommendation system to reflect user's interest.

show abstract

A Community-Based Recommendation System to Reveal Unexpected Interests

Cited by 38 publications

References 15 publications

A Dynamic Community-Based Personalization for e-Government Services

A Dynamic Community-Based Personalization for e-Government Services

An investigation on the serendipity problem in recommender systems

Sclable Collaborative Virtual Environment Considering User's Interests Based on P2P Overlay Network

Contact Info

Product

Resources

About