Student Dropout Prediction in High Education, Using Machine Learning and Deep Learning Models: Case of Ecuadorian University

This work addresses the problem of student dropout in the university environment, where students abandon their studies at any phase of their career, for reasons ranging from personal and economic factors to academic dissatisfaction. This results in the student not starting the degree, or dropping out in early cycles and some cases in the final cycles, affecting the continuity of the study program, and at the same time the students do not reach their profession. With the help of Artificial Intelligence, it is possible to find patterns that allow us to avoid these scenarios early, providing universities with useful information to establish mitigation strategies. This research proposes a prediction model for student dropout in critical subjects, using classic Machine Learning techniques and advanced Deep Learning algorithms, integrated under the KDD method, for analysis of the academic information of the Computer Science degree over 2 years. The results of the model allow us to identify the techniques and algorithms with greater precision and at the same time recognize patterns that establish the dropout trend in core and pre-professional subjects.