基于社会化感知数据多层次学习的服务推荐

项目名称： 基于社会化感知数据多层次学习的服务推荐

项目编号： No.61472253

项目类型： 面上项目

立项/批准年度： 2015

项目学科： 自动化技术、计算机技术

项目作者： 曹健

作者单位： 上海交通大学

项目金额： 80万元

中文摘要： 随着互联网上服务形态的日益多样化、服务数量的快速增长和服务的广泛使用，如何为用户推荐合适的服务已经越来越成为一个迫切需要研究的问题。由于服务提供是一个交互过程，一项服务的达成及其效果不仅取决于自身，也取决于使用它的用户、与它协作的服务和应用的场合，因而服务的准确推荐有赖于对服务、用户及其关系的深入认知。随着服务的广泛使用，服务、服务提供者、用户、服务应用之间已经形成了一个动态变化的、复杂的关系网络，围绕服务的各种社会化的感知数据依托此网络不断产生并在互联网上得到传播。本课题中通过采集和处理服务的社会化感知数据，提出基于这些数据对服务和用户模型进行多层次学习的方法，以设计出能有效利用多源信息、针对用户服务请求行为、适应服务异构性和动态性的服务推荐算法，从而能够为不同模式的应用给出相应的服务推荐方案。

中文关键词： 服务计算；服务推荐；社会化感知数据；多层次学习

英文摘要： With diversification of service types, rapid increment of service amount and wide applications of services in the Internet, how to recommend appropriate services to users becomes an important research problem. Because service provision is essentially an interactive process, the fulfillment and its effects do not only rely on itself but also depend on its users, its collaborative services and its invocation environment. Therefore, the precise recommendation of services should be based on deep understandings of services, users and their relationships. With the wide applications of services, services, service providers, users and applications have formed a dynamic changing and complex relational network. All social sensory data is produced and transmitted across the network. In this project, the approach to collect and process service related sensory data is proposed at first. Then the multilevel learning method of understanding services and users based on this data is studied. Based on this method, the service recommendation algorithms that can make use of multi-source information, accommodate users' service behaviors and consider services' heterogeneous and dynamic features are designed so that we can provide corresponding recommendation solutions to different applications.

英文关键词： Service Computing;Service Recommendation;Social Sensory Data;Multilevel Learning