Predicting early hospital readmission rates for patients with Diabetes

• Used UCI Machine Learning Repository’s Diabetes 130-Hospital Dataset to find the best fitting model for predicting early hospital admission rates in Diabetic patients
• Performed feature engineering steps such as removing unimportant features, replacing and grouping feature values, one hot encoding categorical features and rescaling numerical features
• Employed SMOTE to tackle class imbalance in the target feature
• Performed Grid Search on the following models: Logistic Regression, Random Forest, and Neural Network to select the best model using Cross Validation