摘要
针对现有信任度量方法不能解决实体间评价尺度的差异性而普遍存在信任度量准确性下降的问题,提出了一种基于模糊修正的信任度量算法。通过模糊成员函数表达了信任值的语义,并详细描述了一种推荐信任值的模糊修正算法,在此基础上提出了基于相似度的推荐信任聚合机制。实验结果表明,与PageRank和TidalTrust算法相比,该算法增强了信任度量的准确性,降低了信任决策中的误判率。
In order to deal with the problem of the decline of accuracy of trust metric,which is met due to the ignorance of the difference of rating scale,we presented a new trust metric based on fuzzy adjustment.Firstly,we described the semantic of trust rating and rating scale based on fuzzy member functions.On this basis,we proposed a trust rating adjust method.Lastly,we applied the weigh calculated from similarity matrix to trust aggregation.Analysis and experiment show that compared with Pagerank algorithm and TidalTrust algorithm,the method proposed in this paper has more remarkable enhancements in the accuracy of trust metric and is effective to reduce the false positive and negative in trust decision.
出处
《计算机科学》
CSCD
北大核心
2011年第B10期83-86,共4页
Computer Science
基金
高等学校博士学科点专项科研基金(20093219120024)
自主科研专项基金(2010GJPY056)资助
关键词
评价尺度
信任值修正
信任聚合
信任度量
Rating scale
Trust value adjustment
Trust aggregation
Trust metric
