Analyzing Big Data Using Machine Learning to Identify Factors Affecting College Students Dropout Risk

This article is a comparative analysis of two machine learning methods – the decision tree model and gradient boosting – to predict the risk of college dropout based on data on their demographic characteristics, educational background and socio-economic status. The authors of the study evaluated the accuracy, precision, recall and F1 scores of both models based on test data, identified significant factors affecting the risk of expulsion, and offered recommendations for educational institutions to reduce this risk. During the analysis, social factors were also taken into account, such as the level of student engagement in the learning process and the presence of financial difficulties. Special attention was given to assessing the impact of parents' educational background on students' academic performance. The results obtained may be useful for developing new support strategies for students aimed at increasing the completion rates of their education.