CSAS: Course Schedule Advisory Expert System

Abstract

Student academic advising is an educational decision support process required nowadays to be automated for both the student and advisor to speed up the planning process of course registration by finding out the optimal sequencing or prioritization of students degree requirements to enroll in for next semesters based on their academic standing in a way that suits students’ interests and meets overall graduation requirements within a time frame. This thesis integrated all the rules and constraints from all precedent references relevant to this subject that impacts the scheduling process and proposed a hybrid of model-driven and knowledge-driven decision support system (DSS). Model-driven DSS is based on interoperability between JESS inference engine and Java engine whereas knowledge-driven DSS is based on discovery of new knowledge or patterns using data mining techniques. The solution is named Course Schedule Advisory Expert System (CSAS) implemented as a Web-based application as well as mobile application deployed on android smart device. CSAS analyses data in knowledge systems and allows students to seek quick responses to their queries regarding their plan of study and progress in the program. Technically, inference engine is a hybrid of rule-based engine using JESS (Java Expert System Shell) and knowledge-driven (data mining) based engine using Rapid Miner-java libraries. JESS uses Rete algorithmwhich processes rules and facts in its working memory to generate feasible course registration plan for next semesters of the uncompleted requirements according to contract sheets of each major. JESS version 8 is a Java–based rule engine and scripting environment that supports android platform unlike precedent versions that supported only Java platform (NetBeans). JESS allows dynamic knowledge management in real time. Application retrieves student information from SQL Server knowledge base database via web service application deployed on a communication server for further processing by computational model before rendering results in the GUI component of implemented application whether it’s Web-based or mobile-app. The results of the developed prototype revealed that the model generated accurate results according to system specifications and implemented rules.