基于本体内容相似度的电子商务推荐信任模型

A Recommendation Trust Model Based on Trade Goods Domain Ontology Content Similarity

电子商务环境下,交易双方由于缺乏相互信任的基础和知识,因而交易面临较高的风险.提出一种基于领域本体交易内容相似度,并同时区别熟人节点和陌生人节点推荐可信程度的推荐信任模型(简称DOCSRTrust),给出了一种基于类别的分层次交易商品领域本体构建法,设计了DOCSRTrust的数学表述模型和实现算法.DOCSRTrust模型消除了现有全局信任模型是基于信任值高的节点其推荐也更值得信赖的主观假设,这种主观假设在目前大规模分布式网络环境下并不符合实际情况,因而其客观性和可靠性难以保证.相比之下,分析和仿真实验证明了DOCSRTrust模型更符合当前新型网络应用环境,其在抵抗恶意节点诋毁、遏制协同作弊等较广泛的安全问题上成功率都有较大程度的改善和提高.

Under E-commence environment,because both parties of transaction lack trust foundation and knowledge,transaction face the higher risk.A new recommendation trust model is brought up by the paper,which based on domain ontology transaction content similarity and trust degree of recommendation of distinguished acquaintance nodes from stranger nodes,（abbreviation DOCSRTrust）,simultaneously,and it is also brought up a hierarchical building method of domain ontology transaction goods attribute-oriented,and Mathematical formulation and algorithm of DOCSRTrust model is given by the paper.DOCSRTrust model eliminated a subjective hypothesis that recommendation of a trustworthy node is the more worthy trust in previous global trust models,but in fact,the subjective hypothesis is not to conform to actualities in the current large-scale distributed network environment,so it is the hard to ensure its objectivity and reliability.By contrast,and analyzed and proved the DOCSRTrust is to conform to the current new network application environment by simulation experiment,and there is a greater degree of improvement and enhancement in transactions success ratio that DOCSRTrust holded out malicious nodes＇ slander and hold back malicious node＇s collaborative cheating.