Discovery of Learners' Academic Patterns from a Multi-dimensional Educational Dataset: A Case Study of the Punjab School Education Board, India

Educational Data Mining is an emerging trend in learning management, aimed at identifying changes within the educational system and uncovering hidden factors that influence learners' academic performance. This study utilizes six consecutive years of student examination data, comprising approximately 2 million records from the Punjab School Education Board, India, organized in a database format to identify dependencies among learner attributes. To effectively analyze students' academic performance, this paper proposes a framework combining frequent pattern mining and expert rule mining. It refines the frequent pattern mining algorithm and generates new rules to uncover significant patterns among various attributes in the dataset. Additionally, validation checks are conducted on the resulting pattern set to identify the optimal hypotheses for improving learner academic performance.