Dynamically re-definable ontology based user intension awareness

Abstract

Context awareness has become an important issue in recent years. However, we still do not have a formal mechanism for solving related problems due to the complexity caused by personalization, service diversity, user intension, and so on. In this paper, we investigate the possibility of solving the problems using re-definable ontology. As a concrete example, we study smart phone based automatic service recommendation. The main idea of this work is to record the user behaviors, estimate the user intention based on the recorded data, induce first-order predicate logic rules, and re-define the structure of an ontology system. The re-definable ontology system can become smarter and smarter to understand the user intention and the correlation between users, and recommend proper services in proper timing. Based on the proposed mechanism, a simple prototype system is implemented. This paper provides some preliminary test results, and points out problems for future study.