ABSTRACT
University dropout has increased significantly in Peru before and even more so after the COVID-19 pandemic, which is why public universities need to identify and implement programs to reduce it. The purpose of the work is to determine the Machine Learning algorithm that has the best performance to detect university dropout. This analysis was based on the study of university dropouts in Peru between 2018 and 2021. The population is made up of 652 students, 30% were used for training data and 70% for test data from a data set of 106 valid data, for the development of the classification models, the Anaconda Python language was used through its different libraries, the type of research is applied and descriptive design. It was obtained as a result that the K-Nearest-Neighbor algorithm with an accuracy of 0.91, has better performance to predict university dropout with the academic and socioeconomic variables of the students. In conclusion, the model obtained can help predict, in the first cycles of studies, the students most likely to drop out of their studies, as well as alert the welfare office, the need and attention of individual and group tutoring. © 2022. Revista de Ciencias Sociales. All Rights Reserved.