摘要
为解决Web服务组合过程中服务发现的效率和用户满意度问题,提出了基于情境的Web服务推荐方法。构建情境模型来刻画用户和服务的信息,引入服务缓存机制,利用K-MEANS算法对情境相似的服务进行聚类,实现了服务功能和QoS匹配的初步筛选。利用服务聚类集,综合用户特征和用户评价相似性进行用户情境聚类,并据此对初步结果进行过滤,实现了满足用户个性化需求的优化推荐。实例分析和实验结果表明,该方法具有可行性,并在服务推荐的准确性和时间效率上优于其它方法。
To deal with the efficiency and user satisfaction of service discovery in the process of Web service composition, a met- hod of Web service recommendation based on context is proposed. Firstly, the context model for describing user and service in- formation is built. Secondly, initial screening of service on the function and quality of service is achieved by introducing the ser- vice cache mechanism and using K-MEANS algorithm to cluster services with similarity context. Then the service clustering is exploited, the user profiles and user evaluation similarity are used to cluster user context, and preliminary result is filtered accordingly. Optimizing results, which satisfies the individuality demand, is achieved. The case study and experiment show that the proposed method is feasibility, and is better than other methods in the accuracy and time efficiency of service recommendation.
出处
《计算机工程与设计》
CSCD
北大核心
2014年第3期1115-1120,F0003,共7页
Computer Engineering and Design
关键词
情境
聚类
服务推荐
用户满意度
服务缓存
context
clustering
service recommendation
user satisfaction
service cache
作者简介
古凌岚（1965-），女，广东梅州人，硕士，副教授，研究方向为分布式计算、Web服务、推荐系统。E-mail：Li-Lace@126．com
