December 2024
With a 5-person team, this project applies various machine learning techniques to 500,000+ data points to identify key predictors and risk factors that can inform policy decisions. We engineered features, performed classification analysis, and constructed correlation matrices and heatmaps to explore the relationships between our features. Our modeling comprised logistic regression with backwards elimination, linear and quadratic discriminant analysis, and decision trees to predict whether a certain customer would be interested in auto insurance.