Predicting Student Dropout Rates in Massive Open Online Courses Using an Attention-Based GCNN Model

There is a treasure trove of information, insights, and monitoring possibilities in the mountains of student data that colleges collect. Economic growth, employment prospects, competitiveness, and productivity are all hindered by educational failures and school dropout rates, which impact not only higher education institutions but also students, families, and society at large. A systematic technique consisting of three essential steps is followed in this work. The first step is to collect the dataset, and then for each data point, preprocessing methods are put up. Here, we focus on variables like gender, previous academic performance, and sessional marks, and we transform and handle missing values as part of the data selection process. The importance of the feature scores is then determined using a feature selection procedure. Lastly, an Attention-based GCNN method outperforms more traditional methods like GCN and ACNN while training the model.