Predictive Sepsis Detection Project

This repository contains the code and resources for a Predictive Sepsis Detection project. The project aims to develop a system for automated medical diagnosis using Electronic Health Records (EHR) to predict the onset of sepsis.

Overview

Sepsis is a life-threatening condition that arises when the body's response to infection causes injury to its tissues and organs. Early detection and intervention are critical for improving patient outcomes. This project utilizes machine learning and deep learning algorithms to analyze EHR data and predict the likelihood of sepsis onset.

Contents

• Flask_App: Contains the Flask web application for deploying the predictive model.
  • app.py: Flask application script.
  • Dockerfile: Docker configuration file.
  • random_forest_classifier.pkl: Pre-trained Random Forest classifier model.
  • requirements.txt: List of Python dependencies.
  • static: Directory for static files used in the web application.
  • templates: Directory for HTML templates used in the web application.
• Notebook_for_Predictive_Sepsis_Detection.ipynb: Jupyter Notebook containing the code for data preprocessing, model training, and evaluation.
• Automated Medical Diagnosis: Predictive Sepsis Detection using Electronic Health Records (EHR).pdf: Documentation or research paper detailing the project methodology, results, and deployment.

Supervisor

This project was conducted as an internship at JP Dasolt under the supervision of Vijeth Rai.

Access the Web Application

The web application is hosted on Azure and can be accessed at:

http://medicalcare.azurewebsites.net

Feel free to visit the webpage to interact with the Predictive Sepsis Detection application.