Dynamic Discovery of Personalized Web Services

Web services are a key for today's e-commerce and are gaining an increasing popularity. Because of the availability of extensive variety of services to perform a particular task, it is necessary that users be supported in the ultimate selection of suitable services. For this reason, personalization has become an important aspect in service discovery. Information about service usage is an important component to assess a user's choice and requirements. This information can be gathered from service access logs. This paper presents a logging and web service usage analysis technique in order to provide a means to ensure appropriate personalized services delivery to a client. By using log analysis with web usage mining statistical information is captured. User is thus offered with a set of services that are most interesting to him. Users are clustered into groups according to their past usage of services, and a potential correlation between web services and user groups is determined, using Fuzzy C-Means (FCM) algorithm. These are obtained on the basis of similarity corresponding to the user profiles that contain state information about a client.