How anikesh20/diseases-prediction-system works
Main Components
a) Data Layer:
• Two primary datasets (diabetes.csv and heart.csv)
• A separate storage for user input data (Saved_data from user)
b) Model Training Layer:
• Training notebooks (diabetes.ipynb and heart.ipynb)
• Saved models directory (diabetes_model.sav and heart_disease_model.sav)
c) Application Layer:
• Streamlit web interface (app.py)
• Model inference pipeline
Data Flow
• Dataset → Training Notebooks → Saved Models
• User Input → Saved User Data
• Saved Models → Web Application
Process Flow
• How user interactions flow through the system
• How predictions are generated
• Blue: Data Storage Components
• Green: Processing Components
• Orange: User Interface
• Purple: Model Components
• Rectangles: Main components
• Rounded Rectangles: Processing units
• Hexagons: ML models
Specific Elements
a) Data Layer (Bottom):
[Dataset]
• diabetes.csv
• heart.csv
[User Data Storage]
• Saved_data from user
b) Model Layer (Middle):
[Training Pipeline]
• Training Notebooks
• Saved Models
[Inference Pipeline]
• Model Loading & Prediction
c) Application Layer (Top):
[Streamlit Web Interface]
• User Input Forms
• Prediction Display
Key Aspects to Emphasize
• Clear separation between training and inference paths
• Data persistence points
• Model versioning
• User interaction flows
Technical Details to Note
• Streamlit as the web framework
• Scikit-learn (implied by .sav files) for model training
• CSV-based data storage
• Python-based implementation