Applying Machine Learning Strategies to Educational Outcomes: Modeling Usage on Chicago Public Schools
This summary proposes possible models and applicable considerations for a model-based approach to understanding graduation rates within Chicago’s Public Schools High Schools.
Using a classification model based on a graduation rate threshold, central numerical and categorical features are cleaned, imputed, and standardized, then run through several model types, and checked for accuracy, precision, and F1 scores from their results.
Models with the highest overall performance are then analyzed for feature importance, and re-modeled by categorizing types of features within the original. While exploratory, this paper and study conclude by specifying the importance of different factors within a school, and their effects on schools meeting graduation targets.