Exploring the Procrastination of College Students: A Data-Driven Behavioral Perspective

Procrastination refers to the practice of putting off impending tasks due to the habitual carelessness or laziness. The understanding of procrastination plays an important role in educational psychology, which can help track and evaluate the comprehensive quality of students. However, traditional methods for procrastination analysis largely rely on the knowledge and experiences from domain experts. Fortunately, with the rapid development of college information systems, a large amount of student behavior records are captured, which enables us to analyze the behaviors of students in a quantitative way. To this end, in this paper, we provide a data-driven study from a behavioral perspective to understand the procrastination of college students. Specifically, we propose an unsupervised approach to quantitatively estimate the procrastination level of students by the analysis of their borrowing records in library. Along this line, we first propose a naive ReadingProcrastination (naive RP) model, which considers the behavioral similarity between students for procrastination discovery. Furthermore, to improve the discovery performance, we develop a dynamic ReadingProcrastination (dynamic RP) model by integrating more comprehensive characteristics of student behaviors, such as semester-awareness and month-regularity. Finally, we conduct extensive experiments on several real-world data sets. The experimental results clearly demonstrate the effectiveness of our approach, and verify several key findings from psychological fields.