Files
Abstract
In this thesis, I document practical and theoretical issues concerning the development of IMMGNOSIS, a stateless Web-based expert system that reasons over matters involving U.S. immigration law. I focus on the knowledge engineering aspects of the IMMGNOSIS project, detailing knowledge acquisition, knowledge representation, and inference. Additionally, I present a modified expert-system shell that efficiently handles multiple consultations in a stateless, Web-based environment while relying on only a single instance of the inference application. I evaluate the accuracy of the systems diagnoses, the performance of its stateless architecture, and the potential benefit of putting it into practical use. Finally, I present future plans for knowledge-base expansion and intelligent handling of conflict resolution.