| Summary: | 碩士 === 國立中正大學 === 電機工程研究所 === 93 === User intention identification plays an important role in several domains such as Internet searches, and guided-view systems. If a system is capable of identifying the user intention, it is more convenient and user-friendly for the user to use it. Web services have become one of the hottest topics in IT business in recently years. Web services can range from simple stock quotes or weather forecasts to complex business processes. However, using these services efficiently is not an easy task for the user who is not an expert in computers. In this thesis, we propose our system, called the intention extractor, to extract the user intention and integrate with the Web services agent to satisfy the user’s request for Web services. We propose the intention extractor architecture to implement the intention extraction process. The input string is parsed into a set of action quadruples by referring to the lexical dictionary and domain ontology. A goal model is used to represent user intention. A goal generator uses the action quadruples to generate the related goal models by the goal hierarchy. A goal hierarchy is constructed to describe the scenario that the system interacts with Web services and used to help identify goal models. A goal selector ranks the candidate goal models by user profile and shows them to the user for selecting one. The selected goal model is translated into a query document and the Web service agent will perform it to satisfy the user’s request.
|
|---|
