Association for Public Policy Analysis & Management

In this paper, I create an algorithm to predict which students are eventually going to drop out of high school using information available in 9^th grade. I show that using a naïve model - as implemented in many schools - leads to poor predictions. In addition to this, I explain how schools can obtain more precise prediction by exploiting the big data available to them, as well as more sophisticated quantitative techniques. I also compare the performances of econometric techniques such as Logistic Regression with Machine Learning tools such as Support Vector Machine and LASSO. Finally, I show how budget constraints and policy goals can be taken into account when tuning the model parameters.