Paper Title
30-Day Hospital Readmission Prediction Model for Diabetic Patients within the 30-70 Age Group

Abstract
This study evaluated the performance of four models(Generalized Linear Model (GLM),Under Bagging, Adaboost.M2, and Generalized Boosted Regression Model (GBM)) that were built specifically to identify diabetic patients within the30-70 age group who are at risk for hospital readmission within 30 days. The dataset used in this study was a portion of the dataset from UCI Machine learning website representing 10 years (1999-2008) of clinical care at 130 US hospitals and integrated delivery networks. The resulting derived dataset exhibited class imbalance (11% minority). Feature selection was done using Boruta algorithm. SMOTE was also used to resolve the class imbalance problem. Performance metrics sensitive to imbalanced data were used to evaluate the best model. The resulting readmission prediction model workflow consisted of the following steps: data acquisition, analysis, and preprocessing, data partitioning, feature selection, model building, performance evaluation, model selection, and model tuning. Keywords - Prediction Model, Hospital Readmission, Generalized Linear Model (Glm), Underbagging,Adaboost.M2, Gradient Boosting Machines (Gbm), Boruta, Smote, Imbalanced Datasets