Semester of Graduation
Information Systems and Business Analytics
First Major Professor
Anthony M Townsend
Master of Science (MS)
We live in an era where machine learning and data science play a pivotal role in almost all of the fields. Healthcare is one such field where the implementation of cutting-edge machine learning tools are used to predict, prevent, and cure diseases in a timely manner. Readmission of patients after their discharge from a medical facility has a significant impact on the cost and patient health. In this scenario, this project ventures out to utilize the historic data of diabetes patients to predict their re-admission based on a variety of diagnostic tests performed over the course of the time that the patient is in the hospital. The methodology is to employ machine learning classification algorithms such as Logistic Regression, Decision Trees, Random Forest, K-Nearest Neighbors (KNN), Linear Discriminant Analysis, and Stochastic Gradient Descent to classify a patient as to whether he/she would be readmitted or not. This project uses Recursive Feature Elimination technique to figure out the most important features that can be used as predictors to predict the readmission of patients. This information could be utilized on new patients such that based on the few diagnostic test results performed on the patient while he/she is treated in the hospital, we would be able to get a clearer picture of the patient concerning re-admissions. The model evaluation metrics that were used are Training Accuracy, Testing Accuracy, Precision, Recall, F-1 score, and Confusion Matrix.
Sridhar, Kishor Kumar
Embargo Period (admin only)
Sridhar, Kishor Kumar, "Using Machine Learning to Predict Readmissions of Diabetes Patients" (2020). Creative Components. 683.