נושא הפרוייקט
מספר פרוייקט
מחלקה
שמות סטודנטים
אימייל
שמות מנחים
ניתוח התנהגות סטודנטים בקורס מקוון בעזרת כריית תהליכים
Analysis of students' behavior in an online course with process mining
תקציר בעיברית
תקציר באנגלית
One of the prerequisites for STEM studies in Israel is acquiring mathematical skills. Since 2016, to reduce the level of dropouts and abandonment in the first year, Ben-Gurion University of the Negev offers the online course "Mathematics Proficiency" which allows new students to take an independent online course that provides guidance on the level of mathematics required for them. The purpose of this study is to analyze the course's data, and the student's behavior during this course, and to examine its effect on their achievements in their studies, in the faculties of engineering and natural sciences. For this purpose, we will use a novel technique called Process Mining and statistics for hypothesis testing. The main research method on which the project is based is process mining. This method is suitable for investigating the "Mathematics Proficiency" course because with its help we can identify behaviour patterns of the students and their real learning methods. We would like to find out which tests the students choose to take, in what order, and how long it takes. After using process mining, we used R to perform regression models and find relationships between explanatory variables such as student data and their behaviour patterns in the "Mathematical Proficiency" course, and their grade point average at the end of the studies. With Celonis, we can see the process of each variant, the 14 most common variants (out of 1,167) that contain 71% of the students (out of 5,847 students) do not go through any activity of the Mathematics Proficiency course, but only in the exams. We found the most and least common tests of the course, the mean and median time of the process, the distribution of the course by gender, department, studied year, etc. Following the process mining in Celonis and the understanding of the behavioral patterns, we clustered the students based on both their progress in the Mathematics Proficiency course and their grades in the course tests. We added it as a variable to a linear regression model we built. In these findings, clustering the students entered the model as a significant variable, the percentage of explained variance increased significantly, and in addition, the P-value of the entire model is significant. This suggests that students' engagement and performance in the course have a significant impact on their academic achievements.