An academic advisory system : integrating an academic database management system, an expert system, and a case-based reasoning system

Abstract

Academic advising is a repetitive, time-consuming, and tedious job. In providing academic recommendations, an academic advisor applies three forms of knowledge i.e. formal knowledge, and Formal knowledge is the knowledge about the curriculum as given in the university and department catalogs. The expert knowledge is gained by training and experience. The case-based knowledge is obtained from the accumulated academic cases in the department. The best choice to eliminate the disadvantages of manual academic advising is to automate it. The key to automation lies in the effective application of the computer science field called artificial intelligence (AI).It was therefore considered by the Mechanical Engineering Department at the University of Queensland that a new framework for an Academic Advisory System (AAS) should be developed as the topic of this PhD. This thesis also develops and demonstrates a prototype based on the new framework. The AAS is an integrated system consisting of an Academic Database Management System (ADBMS), an Expert System (ES) based on decision tables, and a Case Based Reasoning System (CBRS). The ES deals with adviser's formal knowledge and expert knowledge, while the CBRS deals with adviser's case-based knowledge. The ADBMS is used to manage large and diverse academic data such as schedules, time tables, and hundreds of students' records and also to prepare input data for the ES and the CBRS.A new framework of the development of a Case Based Reasoning System (CBRS) using the relational model and extending the case based reasoning techniques by combining CBR with GT coding system, fuzzy sets, and rule base has been developed. GT coding system plays the main role as the foundation of the knowledge representation. Fuzzy sets contribute by making the case retrieval more flexible; and rules provide an easy strategy for case adaptation.To test the approach, prototypes of the AAS and the CBRS have been developed. They are PC based and run under Microsoft Windows and were developed using the relational database packages Paradox for Windows and Microsoft Access. The prototypes demonstrates the general techniques needed for academic advising and has been tested to the specific application of advice within the Department of Mechanical Engineering, at The University of Queensland, Australia and the Department of Mechanical Engineering, at The University of Brawijaya, Indonesia.