Hospital Readmission With Machine learning : A Comprehensive Survey and Analysis

Hospital readmissions is a significant concern in healthcare, leading to increased costs and poorerpatient outcomes. Predicting which patients are at high risk of readmission allows healthcareproviders to take preventative measures, improving patient care and reducing the strain on medicalresources. In this project, we design a machine learning model to predict hospital readmissionswithin a specific time frame (e.g., 30 days post-discharge) using patient data, medical history, andtreatment information. The goal of this project is to implement and compare three mostly usedmachine learning algorithms: K-Nearest Neighbors (KNN), Decision Tree, and Random Forest.These algorithms will be trained on a dataset consisting of demographic data, clinical measurements,treatment details, and previous readmission history. We will evaluate and compare theirperformance using various metrics, like accuracy, precision, recall, and F1-score, to determinewhich model provides the most accurate predictions. Through this analysis, we aim to identify keyfactors contributing to hospital readmissions and provide healthcare providers with a predictive toolthat can help optimize patient care, reduce readmission rates, and lower healthcare costs. Theproject also explores the interpretability of the models, with a particular focus on Random Forestdue to its ability to provide feature importance, making it a valuable tool for understanding thefactors influencing readmission risks. In conclusion, the use of machine learning techniques inpredicting hospital readmissions has the potential to improve the efficiency of healthcare systemsand provide data-driven insights for better patient management.