Using Machine Learning and Mixed Effect Models to Predict Undergraduate STEM Dropout

Poster Session IV

Keyword: Decision Making

Abstract: We investigated STEM attrition by tracking 1,243 University of Miami students’ emotions, expectations, and academic goals using Ecological Momentary Assessments. Linear mixed-effects modeling revealed prediction errors in exam grades influence performance goals and major changes, while machine learning achieved 63% classification accuracy

• Anthony Navarro (Presenting Author)
• William Villano (Author)
• Aaron Heller (Author)