摘要
为了解决在电子商务系统中评估没有用户评价或者只有少量评价的商品的信任值问题,提出了一种向用户推荐新商品的信任启动模型。属性对比通过把商品申明的属性值与可以求评价信任值的同类商品的实际属性对比后获得信任值;信任传递通过聚合商店内可以求出评价信任值的其它商品的信任值后获得信任值,通过评估评价可信度等措施降低恶意评价的危害,提高评估信任值的准确度。分析和测试结果表示提出的信任启动模型能达到预测目标,与其它方法比较具有明显的优势。
To solve problems of trust evaluation for a commodity without rating or few ratings in e-commerce system, a trust bootstrap- ping model employed to recommend new commodities was presented. Attribute comparison was applied to obtain trust value through comparing the attribute value declared of the commodity to the attributes of other commodities whose rating trust values could be compu- ted. Trust transitivity was used to obtain trust value through aggregating the rating trust values of other commodities in the same shop. The evaluation of the credible degree of rates reduced the damage of malicious rates and promoted the accuracy of trust values. Analysis and experimental results showed that the proposed trust bootstrapping model could achieve the forecast target and the proposed method has obvious advantages compared to other methods.
出处
《四川大学学报(工程科学版)》
EI
CAS
CSCD
北大核心
2013年第6期15-20,共6页
Journal of Sichuan University (Engineering Science Edition)
基金
贵州省科学技术基金资助项目([2013]29)