关键词: 语义Web服务，动态混合QoS，协同过滤推荐，推荐算法，相似性计算

Abstract: Service recommendation is a main problem in the services computing. However, most of the current Web service recommender systems make recommend for functional properties, but consider less in the QoS attributes of Web services, and do not support dynamic changes of QoS attributes. The personalized Web service recommender model based on the semantic Web services takes advantages of dynamic mixing QoS attributes, and introduces the semantic Web technology into Web services. Under the supervision of the QoS monitor, it effectively monitors the QoS properties change of Web services,and dynamically updates the QoS attributes of the Web service. According to the establishment of user interest model, the system recommends some personalized Web services to the user .In addition, based on the collaborative filtering recommendation technology widely used in personalized recommendation system, this paper car- ried out a series of pretreatment filling, and fully considered the different time scores impact on the recommendation, combined the user interest degree with the user score similarity, and through different weights showed their importance degrec,comprehensivcly calculaed the nearest neighbor set of the target user, ultimately generated recommendation to the user u. To some extent, the system can effectively improve the efficiency, accuracy of the service recommended, and meet the needs of the user query.

Key words: Semantic Web services,Dynamic mixed QoS, Collaborative filtering, Recommendation algorithm, Similarity calculation