Development of an Improved Method for Predicting Educational and Academic Performance of Students, Based on Data Mining and Machine Learning

Universities and educational institutions collect and store a huge amount of data, such as personal and educational information of students. The huge growth of electronic data in universities points to the fact that by using data analysis methods, it is possible to achieve desirable results in the fields of education and research. One of the main challenges of the educational environment is the success rate of students. There is the issue of what are the most important characteristics of students to predict their academic progress and which algorithm is more suitable for making this prediction, and if appropriate results are obtained in the analysis of academic progress, how can managers plan better based on it. In this article, all the possible characteristics of students in an educational institution, collection and some data mining algorithms as well as a proposed method have been implemented on the data and the results have been obtained, checked and compared with each other based on the criteria of Accuracy, Recall and Precision. The decision tree showed the lowest accuracy with 0.864 and the proposed method showed the highest accuracy with 0.935. Also, the most important features that are effective in the academic progress of students were identified. By using this prediction, managers can also remove the obstacles and provide the ground for the progress of students.