Breast Cancer Predictor || Live Preview

A machine learning web application that predicts whether a breast mass is benign or malignant based on measurements from cytology lab results.

Features

• Interactive web interface with adjustable measurements
• Real-time predictions
• Visual representation of measurements using radar charts
• Professional medical-grade predictions

Installation

Prerequisites

1. Python 3.8 or higher
2. Git

Setup Instructions

1. Clone the repository:

git clone https://github.com/prathamhanda/BreastCancer-Predictor.git
cd BreastCancer-Predictor

2. Create and activate a virtual environment:

For Windows:

pip install numpy pandas scikit-learn streamlit plotly altair

python -m venv .venv

source .venv/Scripts/activate

python -m pip install streamlit

pip install plotly scikit-learn

python -m streamlit run app/main.py

For Linux/Mac:

python3 -m venv .venv
source .venv/bin/activate

3. Install dependencies:

pip install -r requirements.txt

4. Run the application:

# Make sure you're in the project root directory
python -m streamlit run app/main.py

5. Open your web browser and go to:

http://localhost:8501

Project Structure

• app/main.py: Main application file
• data/data.csv: Dataset for training
• model/: Contains trained model files
• assets/: CSS and other static files
• confusion_matrices/: Contains model performance visualizations

Usage

1. Launch the application using the instructions above
2. Use the sliders in the sidebar to adjust measurements
3. View the radar chart visualization
4. Check the prediction results

Common Issues and Solutions

1. Permission Errors When Creating Virtual Environment

If you see an error like:

Error: [Errno 13] Permission denied: '...\.venv\Scripts\python.exe'

Solutions:

• Close VS Code and any terminals using the .venv directory
• Delete the existing .venv directory: rm -rf .venv
• Run VS Code as administrator
• Create the virtual environment in your user directory instead:

cd %USERPROFILE%
python -m venv breast-cancer-env
breast-cancer-env\Scripts\activate
cd path/to/Breast-Cancer-Predictor

2. Streamlit Command Not Found

If you see:

bash: streamlit: command not found

Solutions:

• Always use python -m streamlit run app/main.py instead of streamlit run app/main.py
• Make sure you're in the project root directory, not in the app directory
• Verify streamlit is installed: pip list | grep streamlit
• Reinstall if needed: pip install streamlit

3. scikit-learn Version Warnings

If you see warnings like:

InconsistentVersionWarning: Trying to unpickle estimator LogisticRegression from version 1.2.2 when using version 1.6.1

Solutions:

• These warnings are expected and won't affect functionality
• If you want to eliminate warnings, install the exact version:

pip install scikit-learn==1.2.2

4. Feature Names Warning

If you see:

UserWarning: X does not have valid feature names, but StandardScaler was fitted with feature names

Solution:

• This is an expected warning and won't affect the model's predictions
• The warning appears because of scikit-learn version differences

5. General Troubleshooting Steps

1. Verify Virtual Environment:

# Should show (.venv) in prompt
echo %VIRTUAL_ENV%

2. Check Installed Packages:

pip list

3. Verify Working Directory:

# Should be in project root
pwd

4. Clean Installation:

pip uninstall -r requirements.txt -y
pip install -r requirements.txt

5. Port Already in Use: If port 8501 is in use, Streamlit will automatically try the next available port (8502, etc.)

Model Performance

The application includes three machine learning models for breast cancer prediction:

• Random Forest (96% accuracy)
• Support Vector Machine (98% accuracy)
• Logistic Regression (97% accuracy)

Confusion matrices and detailed performance metrics for each model are available in the confusion_matrices directory.

Warning

This application is for educational purposes only and should not be used as a substitute for professional medical diagnosis.

Version Notes

• The model was trained using scikit-learn version 1.2.2
• Compatible with newer versions but will show version warnings
• Tested on Python 3.8+ and Windows 10/11