Personalised and Dynamic Trust in Social Networks Export

@article{citeulike:4034445, abstract = {We propose a novel trust metric for social networks that is personalised and dynamic. It allows to compute the indirect trust between two agents in a social network which are not neighbours based on the direct trust between agents that are neighbours. In analogy with some personalised versions of PageRank, this metric is based on the notion of feedback centrality and overcomes some of the limitations of other trust metrics. It takes into account all the paths in the graph, including cycles. In particular, it does not neglect the patterns responsible for the high clustering coefficient that characterises social networks, as some other algorithms do. From a mathematical point of view, the resulting expression of indirect trust in terms of the original, direct trust matrix involves only simple linear algebra. We also propose a way to make trust dynamic over time and we show, by means of analytical approximations and computer simulations, that it converges to the desired behaviour.}, archivePrefix = {arXiv}, author = {Walter, Frank E. and Battiston, Stefano and Schweitzer, Frank}, citeulike-article-id = {4034445}, citeulike-linkout-0 = {http://arxiv.org/abs/0902.1475}, citeulike-linkout-1 = {http://arxiv.org/pdf/0902.1475}, day = {9}, eprint = {0902.1475}, keywords = {dlpaws, social-network, trust}, month = {Feb}, posted-at = {2009-08-07 15:47:21}, priority = {2}, title = {Personalised and Dynamic Trust in Social Networks}, url = {http://arxiv.org/abs/0902.1475}, year = {2009} }

TY - JOUR ID - citeulike:4034445 L3 - citeulike-article-id:4034445 N2 - We propose a novel trust metric for social networks that is personalised and dynamic. It allows to compute the indirect trust between two agents in a social network which are not neighbours based on the direct trust between agents that are neighbours. In analogy with some personalised versions of PageRank, this metric is based on the notion of feedback centrality and overcomes some of the limitations of other trust metrics. It takes into account all the paths in the graph, including cycles. In particular, it does not neglect the patterns responsible for the high clustering coefficient that characterises social networks, as some other algorithms do. From a mathematical point of view, the resulting expression of indirect trust in terms of the original, direct trust matrix involves only simple linear algebra. We also propose a way to make trust dynamic over time and we show, by means of analytical approximations and computer simulations, that it converges to the desired behaviour. TI - Personalised and Dynamic Trust in Social Networks KW - dlpaws KW - social-network KW - trust AU - Walter, Frank AU - Battiston, Stefano AU - Schweitzer, Frank PY - 2009/Feb/09/ UR - http://arxiv.org/abs/0902.1475 ER -