MediX11 - A Multi-Disease Prediction System
Overview
MediX11 is a comprehensive machine learning and deep learning-driven health analysis system that predicts 11 critical diseases with high accuracy. It leverages data science, web scraping, and AI models to provide insights into medical conditions, enabling early diagnosis and better treatment planning.
Features
Multi-Disease Prediction - Covers 11 major diseases, from diabetes to neurological disorders.
Deep Learning & ML Models - Utilizes advanced algorithms for high accuracy.
Early Detection & Survival Analysis - Predicts disease onset and survival rates.
Medical Data Scraping - Extracts real-time research insights from PubMed.
Image-Based & Clinical Data Models - Supports diverse medical datasets.
Deployed Risk Assessment - 10-year CHD (Coronary Heart Disease) risk prediction live on Streamlit Cloud at chdvis.streamlit.com.
Included Notebooks
| Notebook | Description |
|---|---|
| Depression.ipynb | Predicts depression risk based on patient responses. |
| Diabetes.ipynb | ML-based diabetes risk assessment. |
| New_dementia.ipynb & dementia.ipynb | Dementia detection models. |
| Parkinson.ipynb & new_parkinsons.ipynb | Parkinson's disease prediction using patient data. |
| Patient_Survival_Prediction.ipynb | Predicts survival rates for critically ill patients. |
| early_stage_detection.ipynb | Identifies diseases at an early stage for better prognosis. |
| eye_disease_detection.ipynb | Deep learning for detecting eye-related diseases from images. |
| heart_disease_prediction.ipynb | Predicts heart disease risk using patient health metrics. |
| PubMedScrapper_Combined_Summary.ipynb | Collects medical research papers from PubMed using Biopython and summarizes them using a BART model. |
| Cardiovascular_Risk_Prediction.ipynb | Predicts 10-year coronary heart disease (CHD) risk using clinical parameters. |
| 10-Year-CHD-Prediction-App | Deployed Streamlit app for CHD risk assessment, available at chdvis.streamlit.com. |
| SKIMLIT.ipynb | Uses advanced NLP and deep learning techniques to classify biomedical literature into key sections like objectives, methods, and results. |
Technologies Used
Python (Pandas, NumPy, Scikit-learn, TensorFlow)
Web Scraping (BeautifulSoup, Biopython for PubMed data)
NLP (Transformer-based models like BART for summarization)
Deep Learning (CNNs for image-based diagnosis, NLP models for text processing)
Streamlit (for web-based disease risk prediction)
How to Use
- Clone the repository:
git clone https://github.com/your-repo/medix11.git
cd medix11 - Run Jupyter Notebook:
jupyter notebook
- Open the desired
.ipynbfile and run the model. - For CHD Risk Prediction, visit chdvis.streamlit.com and input clinical parameters to get risk assessment.
Future Enhancements
Add more disease prediction models (e.g., cancer, liver diseases).
Integrate real-time patient data from wearables.
Improve accuracy using transformer-based AI models.
Expand Streamlit dashboard with more interactive visualizations and predictions.
Contributors
Vishal Sarup Mathur
Contact: official.vishal.sarup.mathur.1@gmail.com
Aditya Singh
Contact: csaditya038@gmail.com