A cognitive knowledge-based model for an academic adaptive e-advising system

Abstract

Aim/Purpose This study describes a conceptual model, based on the principles of concept algebra that can provide intelligent academic advice using adaptive, knowledge-based feedback. The proposed model advises students based on their traits and academic history. The system aims to deliver adaptive advice to students using historical data from previous and current students. This data-driven approach utilizes a cognitive knowledge-based (CKB) model to update the weights (values that indicate the strength of relationships between concepts) that exist between student's performances and recommended courses. Background A research study conducted at the Public Authority for Applied Education and Training (PAAET), a higher education institution in Kuwait, indicates that students' have positive perceptions of the e-Advising system. Most stu-dents believe that PAAET's e-Advising system is effective because it allows them to check their academic status, provides a clear vision of their academic timeline, and is a convenient, user-friendly, and attractive online service. Stu-dent advising can be a tedious element of academic life but is necessary to fill the gap between student performance and degree requirements. Higher edu-cation institutions have prioritized assisting undecided students with career decisions for decades. An important feature of e-Advising systems is person-alized feedback, where tailored advice is provided based on students' charac-teristics and other external parameters. Previous e-Advising systems provide students with advice without taking into consideration their different attrib-utes and goals. Methodology This research describes a model for an e-Advising system that enables stu-dents to select courses recommended based on their personalities and aca-demic performance. Three algorithms are used to provide students with adaptive course selection advice: The knowledge elicitation algorithm that rep-resents students' personalities and academic information, the knowledge bonding algorithm that combines related concepts or ideas within the knowledge base, and the adaptive e-Advising model that recommends rele-vant courses. The knowledge elicitation algorithm acquires student and aca-demic characteristics from data provided, while the knowledge bonding algo-rithm fuses the newly acquired features with existing information in the data-base. The adaptive e-Advising algorithm provides recommended courses to students based on existing cognitive knowledge to overcome the issues asso-ciated with traditional knowledge representation methods. Contribution The design and implementation of an adaptive e-Advising system are chal-lenging because it relies on both academic and student traits. A model that incorporates the conceptual interaction between the various academic and student-specific components is needed to manage these challenges. While other e-Advising systems provide students with general advice, these earlier models are too rudimentary to take student characteristics (e.g., knowledge level, learning style, performance, demographics) into consideration. For the online systems that have replaced face-to-face academic advising to be effec-tive, they need to take into consideration the dynamic nature of contempo-rary students and academic settings. Findings The proposed algorithms can accommodate a highly diverse student body by providing information tailored to each student. The academic and student el-ements are represented as an Object-Attribute-Relationship (OAR) model. Recommendations for Practitioners The model proposed here provides insight into the potential relationships be-tween students' characteristics and their academic standing. Furthermore, this novel e-Advising system provides large quantities of data and a platform through which to query students, which should enable developing more ef-fective, knowledge-based approaches to academic advising. Recommendation for Researchers The proposed model provides researches with a framework to incorporate various academic and student characteristics to determine the optimal advi-sory factors that affect a student's performance. Impact on Society The proposed model will benefit e-Advising system developers in imple-menting updateable algorithms that can be tested and improved to provide adaptive advice to students. The proposed approach can provide new insight to advisors on possible relationships between student's characteristics and current academic settings. Thus, providing a means to develop new curricu-lums and approaches to learning. Future Research In future studies, the proposed algorithms will be implemented, and the adaptive e-Advising model will be tested on real-world data and then further improved to cater to specific academic settings. The proposed model will benefit e-Advising system developers in implementing updateable algorithms that can be tested and improved to provide adaptive advisory to students. The approach proposed can provide new insight to advisors on possible rela-tionships between student's characteristics and current academic settings. Thus, providing a means to develop new curriculums and approaches to course recommendation.