ONLINE COURSE IN-ACTIVE STUDENTS PREDICTION
Ms.D.Kiruthika1, C.Nakul Anand2, K.Vasanth3, K.Deepak4, S.Shanthosh5
1Associative Professor, Department of Information Technology, Nandha Engineering College-Erode- 638052, Tamilnadu, India.
2,3,4,5UG Scholar, Information Technology, Nandha Engineering College-Erode- 638052, Tamilnadu, India.
E-mail : nakulanandc@gmail.com
Abstract. Virtual Learning Environments (VLEs), Learning Management Systems (LMS), and Massive Open Online Courses (MOOCs) are just a few of the online learning platforms that make it possible for thousands or even millions of students to learn according to their interests and without being restricted by time or space. Online learning platforms have many advantages, but they also face a number of disadvantages, such as students' lack of interest, high dropout rates, low engagement, self-regulation, and being forced to set their own goals.
In this study, we propose a predictive model that looks at the issues that at-risk students face and then makes it easier for teachers to intervene quickly to get students more engaged in their studies and better performing. Various machine learning (ML) and deep learning (DL) algorithms are used to train and test the predictive model to determine how students learn based on their study variables. The accuracy, precision, support, and f-score are used to compare the performance of various ML algorithms.
In the end, the ML algorithm with the highest f-score metric, accuracy, precision, recall, and support is chosen to create the predictive model at various percentages of course length. Instructors can use the predictive model to identify students who are at risk early in the course, allowing for prompt intervention and avoiding student dropout. Our findings demonstrated the significance of time-dependent variables, students' assessment scores, engagement intensity, or click stream data, and online learning.
