| With the development of Internet technology, e-commerce is very popular nowadays. Many traditional industries, such as tourism, are embracing e-commerce. The rapid growth of service-oriented architecture and cloud computing brings users lots of services. As a result, users’ decision-making become more and more complex. Accurate personalized recommendation technology is needed badly to provide users personalized services. The author’s research group has been doing research on service discovery, service description, service composition, service execution, service recommendation and user modeling. The author participated in research on user semantic modeling, one of the key technologies in service composition and recommendation. Semantic user model can be used to identify the user preference and support the services recommend framework.Traditional recommendation systems are mainly focus on single or separate resources. In the tourism field, relationships between services tend to be very complex. Services should not be simply bundled or packaged. For example, attractions have a geographical difference and transport services have geographic and temporal association. So recommendation to the user not only depends on the user’s preference, but also constraints between services in time and space. Therefore, only identifying and using semantic information between services and users can we achieve a reasonable portfolio of services and accurate user recommendation.This study uses travel semantic dictionary to express domain knowledge and design user-based semantic model. The domain semantics dictionary defines related concepts, attributes and constraints in tourism field, playing an important role in understanding of users and travel services. We first build specific semantic dictionary and design semantic dictionary-based user semantic model, then implement prototype system to achieve service composite recommendation and ultimately improve the user experience. |
PDF Full Text Request