w. Yuan et al/ Knowledge-Based Systems 23(2010)232 example of the applications, and show how the small-worldness of MKE/ KEIT. [2009-S-033-01, Development of Saas Platform for S/ the trust network contributes to the applica pecifically, we w Service of Small and Medium sized Enterprises. propose a novel TARS model by using L as MTPD for TARS. The simulation results show that: by involving recommenders that References are within LI hops away from the active user, it is possible to achieve high rating coverage and recommender coverage, while [1] P. Massa, P. Avesani. it is computational exponentially less expensive than using a great- systems, in: Proceedings 20pP+ Conference on the move to er value of MTPD. On the other hand, by using [L] as MTPD, the er- 12 P, 02 ACM Conference on Recommender Systems, Minneapolis, MN, USA, ACM, ommender systems, in: Proceed ror of the predicted ratings is less than the error of using a smaller value of MTPD. These simulation results verify the effectiveness of ur proposed methodology of TARS [31 E. Gray, J. Seigneur, Y. Chen, C. Jensen, Trust propagation in small worlds, in Our future work focuses on several aspects. Firstly, we will mprove the existing TARs models with the implicit trust net [4 M. Venkatraman, B. Yu, M.P. Singh, Trust and reputation management in a york. Existing works of TARS focus on using the explicit trust, while it is sometimes time consuming or expensive to get [51 x Guo, eling small-world trust networks licit trust. Explicit trust refers to the trust that should be explic ternational Symposium on Ubiquitous Multimedia Computing 2008, UMC itly pointed out by the users. These explicit trust statements ar then used as the inputs of TArs with the recommendations to th International Conference on Intelligent User Ints Proceedings of the predict the ratings. Though the explicit trust based TARS models have high rating prediction coverage and high rating prediction [7 P Bedi, H Kaur, S Marwaha, Trust based recommender system for semantic web, in: Proceedings of the 2007 International Joint Conferences on Artificial ccuracy, the explicit trust statements are not always available Intelligence, Hyderabad, India, pp. 2677-2682. Therefore, we will try to use other cheap and easy available trust n: Proceedings of Www-08, 2008, pp. 199-208. TARS. Secondly, we will focus on how to filter out the unfair rec- [9]G. Pitsilis, L Mar shall, A trust-enabled P2P recommender system, in: ommendations for TARS. TARS suggests information to the active sers based on the recommendations given by various Infrastructure for collaborative Enterprises, 2006. D0.59-64ing technologi [10 P. Massa, P. Avesani, Controversial users demand local trust metrics: an menders. However, there may exist some self-interested recom- own gains(perhaps at the cost of others ). So it is essential to 111 P. Massa. P Avesani, Trust metrics in recommender systems, in: Computing avoid or reduce the influence of the unfair positive or negative [12] P. Avesani, P. Massa, R. Tiella, A trust-enhanced recommender system recommendations from the self-interested recommenders for ication: moleskiing, in: Proceedings of the 2005 ACM Symposium this purpose, we intend to introduce the users'distrust state nents into our TARS model. By analyzing the recommendations [ 13] M. Newman, A Barabasi, D.J. Watts. The Structure and Dynamics of Networks first ed, Princeton University Press, given by each users distrusted recommenders and the relation- [14].A small world web. in: Proceedings of the Third European ship between the trust statements and the distrust statements Conference on Research and Advanced Technology for Digital Libraries, the reliable recommendations will be chosen for the rating aggre- [151 H. Ebel, L Mielsch, S. Bornholdt, Scale-free topology of e-mail networks gations of our proposed TARS model. Acknowledgements International Conference on Hybrid Intelligent Systems 2006, HIS06, 2006, p. The authors would like to thank anonymous reviewers and frequency, small-world human brain functional network with highly the editors of the journal for their valuable comments. This re- [181 Ds. 6 )512-527 Bullmore, Small-world brain networks, Neuroscientist 12 search was supported by the MKE(Ministry of Knowledge Econ- my), Korea, under the (Information Technology Research [19] P Mass. Center)support program supervised by the inta(Institute of Infor mation Technology Advancement) (IlTA-2009-(C1090-0902 [20 D Watts, Small Worlds: The Dynamics of Networks Between Order and Randomness, Princeton University Press, 1999. 0002)). Also, this work was supported by the IT R&D program ofexample of the applications, and show how the small-worldness of the trust network contributes to the applications. Specifically, we propose a novel TARS model by using dLe as MTPD for TARS. The simulation results show that: by involving recommenders that are within dLe hops away from the active user, it is possible to achieve high rating coverage and recommender coverage; while it is computational exponentially less expensive than using a great￾er value of MTPD. On the other hand, by using dLe as MTPD, the er￾ror of the predicted ratings is less than the error of using a smaller value of MTPD. These simulation results verify the effectiveness of our proposed methodology of TARS. Our future work focuses on several aspects. Firstly, we will improve the existing TARS models with the implicit trust net￾work. Existing works of TARS focus on using the explicit trust, while it is sometimes time consuming or expensive to get the ex￾plicit trust. Explicit trust refers to the trust that should be explic￾itly pointed out by the users. These explicit trust statements are then used as the inputs of TARS with the recommendations to predict the ratings. Though the explicit trust based TARS models have high rating prediction coverage and high rating prediction accuracy, the explicit trust statements are not always available. Therefore, we will try to use other cheap and easy available trust sensitive information to generate the implicit trust statements for TARS. Secondly, we will focus on how to filter out the unfair rec￾ommendations for TARS. TARS suggests information to the active users based on the recommendations given by various recom￾menders. However, there may exist some self-interested recom￾menders who give unfair recommendations to maximize their own gains (perhaps at the cost of others). So it is essential to avoid or reduce the influence of the unfair positive or negative recommendations from the self-interested recommenders. For this purpose, we intend to introduce the users’ distrust state￾ments into our TARS model. By analyzing the recommendations given by each user’s distrusted recommenders and the relation￾ship between the trust statements and the distrust statements, the reliable recommendations will be chosen for the rating aggre￾gations of our proposed TARS model. Acknowledgements The authors would like to thank the anonymous reviewers and the editors of the journal for their valuable comments. This re￾search was supported by the MKE (Ministry of Knowledge Econ￾omy), Korea, under the ITRC (Information Technology Research Center) support program supervised by the IITA (Institute of Infor￾mation Technology Advancement) (IITA-2009-(C1090-0902- 0002)). Also, this work was supported by the IT R&D program of MKE/KEIT. [2009-S-033-01, Development of SaaS Platform for S/ W Service of Small and Medium sized Enterprises]. References [1] P. Massa, P. Avesani, Trust-aware collaborative filtering for recommender systems, in: Proceedings of Federated International Conference on the Move to Meaningful Internet, 2004, pp. 492–508. [2] P. Massa, P. Avesani, Trust-aware recommender systems, in: Proceedings of the 2007 ACM Conference on Recommender Systems, Minneapolis, MN, USA, ACM, 2007, pp. 17–24. [3] E. Gray, J. Seigneur, Y. Chen, C. Jensen, Trust propagation in small worlds, in: Proceedings of First International Conference on Trust Management (iTrust’03), 2003, pp. 239–254. [4] M. Venkatraman, B. Yu, M.P. Singh, Trust and reputation management in a small-world network, in: Proceedings of Fourth International Conference on MultiAgent Systems, 2000, pp. 449–450. [5] X. Guo, X. Li, Y. Qin, C. Chen, Modeling small-world trust networks, in: International Symposium on Ubiquitous Multimedia Computing 2008, UMC ’08, 2008, pp. 154–159. [6] J. O’Donovan, B. Smyth, Trust in recommender systems, in: Proceedings of the 10th International Conference on Intelligent User Interfaces, San Diego, California, USA, ACM, 2005, pp. 167–174. [7] P. Bedi, H. Kaur, S. Marwaha, Trust based recommender system for semantic web, in: Proceedings of the 2007 International Joint Conferences on Artificial Intelligence, Hyderabad, India, pp. 2677–2682. [8] R. Andersen, C. Borgs, J. Chayes, U. Feige, A. Flaxman, A. Kalai, V. Mirrokni, M. Tennenholtz, Trust-based recommendation systems: an axiomatic approach, in: Proceedings of WWW-08, 2008, pp. 199–208. [9] G. Pitsilis, L. Marshall, A trust-enabled P2P recommender system, in: Proceedings of 15th IEEE International Workshops on Enabling Technologies: Infrastructure for Collaborative Enterprises, 2006, pp. 59–64. [10] P. Massa, P. Avesani, Controversial users demand local trust metrics: an experimental study on epinions.com community, in: Proceedings of 20th national conference on Artificial intelligence, 2005, pp. 121–126. [11] P. Massa, P. Avesani, Trust metrics in recommender systems, in: Computing with Social Trust, 2009, pp. 259–285. [12] P. Avesani, P. Massa, R. Tiella, A trust-enhanced recommender system application: moleskiing, in: Proceedings of the 2005 ACM Symposium on Applied Computing, pp. 1589–1593. [13] M. Newman, A. Barabasi, D.J. Watts, The Structure and Dynamics of Networks, first ed., Princeton University Press, 2006. [14] L.A. Adamic, The small world web, in: Proceedings of the Third European Conference on Research and Advanced Technology for Digital Libraries, Springer-Verlag, 1999, pp. 443–452. [15] H. Ebel, L. Mielsch, S. Bornholdt, Scale-free topology of e-mail networks, Physical Review E 66 (2002). [16] M. Markosova, P. Nather, Language as a small world network, in: Sixth International Conference on Hybrid Intelligent Systems 2006, HIS’06, 2006, p. 37. [17] S. Achard, R. Salvador, B. Whitcher, J. Suckling, E. Bullmore, A resilient low￾frequency, small-world human brain functional network with highly connected association cortical hubs, Journal of Neuroscience 26 (2006) 63–72. [18] D.S. Bassett, E. Bullmore, Small-world brain networks, Neuroscientist 12 (2006) 512–523. [19] P. Massa, K. Souren, Trustlet, open research on trust metrics, in: Proceedings of the 2nd Workshop on Social Aspects of the Web (SAW 2008), 2008, pp. 31–43. [20] D.J. Watts, Small Worlds: The Dynamics of Networks Between Order and Randomness, Princeton University Press, 1999. 238 W. Yuan et al. / Knowledge-Based Systems 23 (2010) 232–238