Predicting Student Academic Achievement Using Data Mining and Deep Learning Techniques in Educational and Medical-Legal Contexts

Abstract

Educators are starting to take an interest in intelligent technology development. Conventional processing methods may be inadequate and skewed due to the exponential growth of educational data. Therefore, it is more crucial than ever to replicate data mining research technologies for use in the field of education. Using relevant theories of grouping, discriminating, and convolutional neural networks, this study assesses and forecasts students' academic performance in order to avoid inaccurate assessment results and to plan for their future performance. Using a statistic that has never been utilized in the K-means approach before to improve the clustering-number determination is the first advice from this study. After that, we'll assess the K-means method's clustering efficacy using discriminant analysis. A convolutional neural network, which can be trained and evaluated with labelled data, is shown. The produced model can be used to forecast future performance. The last step is to use two metrics in two cross validation methods to assess the constructed model and verify the predicted findings. The experimental results show that the statistic makes the results more predictable and answers the quantitative and objective question of how to get the clustering number using the K-means method. The accuracy of 98.34%, recall of 97.85%