In this project, we constructed an ML pipeline that predicted changes in the overall severity of depression symptoms in adolescents with high symptoms over a three-month period. We computed the residuals from a linear regression model, predicting the 3-month depression sum score based on the baseline depression sum score.