With the development of big data and machine learning technologies in recent years, the results of disease prediction may become more accurate. Doctors are working with statisticians and computer scientists to develop better tools to predict disease. Experts in this field are studying methods to determine, develop, and fine-tune machine learning algorithms and models to provide accurate predictions. In this project, we hope to use data collected from patient demographics, medical health records, and other aspects to develop a robust machine learning model. The goal of our model is to retrieve similar patients and recommend early stage treatment for prevention.
Dataset is from MIMIC-III Critical Care Database. MIMIC-III (Medical Information Mart for Intensive Care III) is a freely-accessible database, and it includes health data between 2001 and 2012 from over 40,000 ICU patients of the Beth Israel Deaconess Medical Center. Submitting a request to access the dataset is required.
Dataset is available from https://mimic.physionet.org/about/mimic/ .
Implementation of visualization is built with Shiny and R.