An Active Object-Oriented Approach for Human Information System Retrieval

摘要

This paper focuses on issues related to the understanding of human-computer interaction, in the form of Natural Language (ML) interface for object-oriented application framework. The Lexical-Functional Grammar (LFG) theory has been adopted to constitute the theoretical syntactic, and semantic presentations of this study. Answers to raised query not only based on direct stored text, but on understanding of text within the domain framework. The work has been conducted by developing a general Object Model for traffic accident domain. This Object Model is used to capture the functional behaviour of verbs and their linguistic associations in a given query. The combination of object's behaviour in the Domain Model with the LFG Structures rules, resolved most, if not all, of the common ambiguities found in NL for that domain and enhances the understanding ability of the proposed framework. The prototype has been implemented using Object Management Workbench tools and Kappa Rule Based System, running under a Sun/Solaris platform. It has been built by experimenting with various newspaper stories. The prototype has been tested by a set of queries and produced the expected results.