Process mining application for discovering student learning paths

Personalized learning has been gaining attention in the research area of online learning systems to enhance learning process efficiency. Several studies have been successfully used educational data mining to discover insight knowledge to improve personalized learning. However, educational data mining is a data-centric analysis, the visual presentation of the learning process from the start to the end is still absent as well as the learning process variants. The objective of this study is to enhance personalized learning by extracting learning process models from student groups who have different learning backgrounds. We proposed a method to analyze event logs from real-world educational data which has a high degree of dynamic and complexity. The process mining is the basis for the purposed method. The interquartile range method is applied to eliminate outliers from event logs. And heuristics miner is used for extracting the learning process models. The result of the proposed method provides visual presentations of actual learning process models and variants from different groups of students.