针对现有服务选择中服务推荐技术的不足,提出一种基于偏好推荐的服务选择(trustworthy services selection based on preference recommendation,简称TSSPR)方法.首先搜索一组偏好相似的推荐用户,并通过皮尔逊相关系数计算用户的评价相似度,然后基于用户的推荐等级、领域相关度和评价相似度等对用户的推荐信息进行过滤,从而使推荐信息更为可信.模拟实验结果表明,通过正确的参数设置,该方法能够有效地解决推荐算法中冷启动、拒葬信息不缝确等问颗.
This paper presents a Trustworthy Services Selection Based on Preference Recommendation (TSSPR) method that assists users in selecting the right Web services, according to their own preferences. First, a group of recommenders that have similar preferences are found, and then the similarity rating is computed by using the Pearson correlation method. Second, filtering services based on the user's recommending level, relative domain degrees, and similarity ratings can improve the quality of recommendations. Experimental results show that given an appropriate setting, this method can effectively solve the weaknesses of recommendation systems, such as sparseness, cold starts, and inaccurate recommendations.