Intention Interpretation for Service Request Using User Models

Abstract

A method to extract user intention from the service request string entered by the user is proposed and applied in semantic Web service systems. Imprecise user intentions are modeled with computable goal models. Three kinds of background knowledge are used when extracting the user intention: domain ontology, goal structure, and user profiles. Domain ontology is constructed to represent the concepts in the application domain, the goal structure indicates the system capabilities, and the user profiles are for the intelligent system to infer the user's preference. The input terms will be parsed into different word sense sets. With the domain ontology, possible senses of the terms will be identified. The goal structure is used for mapping the user's requests to system capabilities. The user models will be used to aid the selection of a personalized candidate goal model for representing the user intention. The system can base on the selected goal model to generate and perform a series of actions for achieving the goal model.