A systematic review on machine learning models for online learning and examination systems

Area of Science:

• Computer Science
• Education Technology
• Artificial Intelligence

Background:

• The COVID-19 pandemic disrupted traditional education, necessitating remote learning and assessment solutions.
• Machine learning (ML) emerged as a critical technology for facilitating online education and examinations during lockdowns.
• Existing literature lacked a comprehensive review of ML's role in managing examinations during this period.

Purpose of the Study:

• To systematically review the application of Machine learning (ML) in Lockdown Exam Management Systems.
• To analyze the significance of ML across the entire examination lifecycle: pre-exam, during, and post-examination.
• To identify and categorize ML algorithms used in various examination processes.

Main Methods:

• A systematic review methodology was employed.
• 135 studies published within the last five years were evaluated.
• Analysis focused on ML algorithms (supervised and unsupervised) applied to exam management.

Main Results:

• Machine learning (ML) significantly supports exam processes including authentication, scheduling, proctoring, and fraud detection.
• ML aids in predicting at-risk students, enabling adaptive learning, and monitoring student progress.
• Both supervised and unsupervised ML algorithms were identified and categorized for different exam stages.

Conclusions:

• Machine learning (ML) is integral to the digital transformation of examinations, offering solutions for remote assessment.
• The review highlights the challenges and proposes solutions for implementing ML in exam management systems.
• Further research is needed to address the complexities and ethical considerations of ML in educational assessments.