🩺 HealthNER

HealthNER is a Natural Language Processing (NLP) project built using spaCy to automatically extract meaningful medical entities such as diseases, symptoms, and drugs from unstructured clinical text. This allows healthcare data to be converted into a more structured and analyzable format.

🌐 Live Demo

Try it now: https://healthner.streamlit.app/

---

🚀 Key Features

• 🔍 Extracts Diseases, Symptoms, and other entities from raw medical text
• 🧠 Trains a custom Named Entity Recognition (NER) model using the NCBI Disease dataset
• 📈 Evaluates model performance using Precision, Recall, and F1 Score
• 🖼️ Visualizes entities using spaCy's displacy and Streamlit UI
• 📦 Deployable as an interactive web app
• 🌙 Dark/Light mode toggle for better user experience
• 📊 Interactive charts and analytics with entity distribution
• 📄 Export functionality (PDF/CSV) for analysis reports
• 🕓 Analysis history tracking

---

🛠️ Tech Stack

• Python - Core programming language
• spaCy - NLP library for entity recognition
• Streamlit - Web app framework
• Plotly - Interactive visualizations
• Pandas - Data manipulation
• FPDF - PDF report generation

---

📊 Screenshots

💻 Streamlit Interface

📥 Input

📤 Output

---

📂 Project Structure

.
├── convert_ncbi_to_spacy.py     # Converts NCBI dataset into spaCy format
├── train.py                     # Trains the NER model using spaCy
├── test_model.py                # Tests the model on simple input
├── test_model_on_data.py        # Tests on complex clinical paragraphs
├── streamlit_app.py             # Streamlit UI for real-time testing
├── visualize.py                 # Saves NER visualizations as HTML
├── requirements.txt             # Python dependencies
├── output/                      # Trained model files
├── README.md                    # Project overview

---

🚀 Getting Started

Prerequisites

• Python 3.7+
• pip package manager

Installation

1. Clone the repository

   git clone https://github.com/DSUCS0018/HeallthNER.git
   cd HeallthNER

2. Install dependencies

   pip install -r requirements.txt

3. Download spaCy English model

   python -m spacy download en_core_web_sm

Usage

Training the Model

# Convert NCBI dataset to spaCy format
python convert_ncbi_to_spacy.py

# Train the custom NER model
python train.py

Testing the Model

# Test on simple input
python test_model.py

# Test on complex clinical data
python test_model_on_data.py

Running the Web App

streamlit run streamlit_app.py

The app will be available at http://localhost:8501

---

📈 Model Performance

The custom NER model is trained on the NCBI Disease dataset and evaluated using standard NLP metrics:

• Precision: Accuracy of entity predictions
• Recall: Coverage of actual entities
• F1 Score: Harmonic mean of precision and recall

---

🎯 Use Cases

• Clinical Research: Extract entities from research papers and clinical notes
• Healthcare Analytics: Convert unstructured patient data into structured format
• Medical Documentation: Automatically tag and categorize medical documents
• Drug Discovery: Identify relationships between diseases, symptoms, and treatments

🙏 Acknowledgments

• spaCy team for the excellent NLP library
• NCBI for providing the disease dataset
• Streamlit for the amazing web app framework

---

📧 Contact

Your Name - [Your Email] - [Your LinkedIn]

Project Link: https://github.com/DSUCS0018/HeallthNER

Live Demo: https://heallthner.streamlit.app/