Detecting Symptoms of Low Performance Using Production Rules.

摘要: E-Learning systems offer students innovative and attractive ways of learning through augmentation or substitution traditional lectures exercises with online material. Such material can be accessed at any time from anywhere using different devices, personalized according to the individual student's needs, goals knowledge. However, authoring evaluation this remains a complex task. While many researchers focus on support, not much has been done facilitate e-Learning applications, which requires processing vast quantity data generated by students. We address problem proposing an approach for detecting potential symptoms low performance in courses. It supports two main steps: generating production rules C4.5 algorithm filtering most representative rules, could indicate In addition, evaluated log files student activity versions Web-based quiz system. paper we try e- Learning based analysis data. Evaluation is time-consuming One challenges instructor evaluating system lack evidence students, since he/she access only their interactions cannot observe behaviors, receive feedback them timely manner. Due fact that behaviors are hidden inside interactions, instructors should analyze order assess evaluate software. Furthermore, generate (log files). These consist records students' actions within Since often very big,