This chapter addresses the classification of at-risk students in educational settings using machine learning approaches in R. Transitioning from regression-based predictions of students’ grades covered by the previous chapter, the focus here shifts to identifying broader categories of academic performance, such as low achievers or potential dropouts. Early identification of such students enables timely interventions, one of the main goals of learning analytics. The process is first illustrated through a Random Forest classifier, using engagement indicators to classify students into high and low achievers. The chapter demonstrates the complete modeling workflow, including data preparation, model training, and evaluation using performance metrics. Additionally, the tidymodels framework is explored as a more modern alternative that enables easy comparison with other AI / machine learning algorithms like Naive Bayes or Support Vector Machine.