Breast cancer is a significant health concern, with early detection being crucial for improving patient outcomes and reducing treatment costs. This project demonstrates the development and evaluation of three machine learning models aimed at accurately predicting whether a breast tumor is benign or malignant. By leveraging these models, healthcare providers can potentially make more informed decisions, leading to earlier interventions and better patient prognosis.

The primary goal of this project is to provide a reliable, automated tool for breast cancer detection that can assist radiologists and oncologists in making quicker and more accurate diagnoses. Implementing this tool could lead to:

Reduced Diagnostic Time: Automating the initial screening process, allowing healthcare professionals to focus on more complex cases.

Improved Accuracy: Enhancing diagnostic precision, reducing the likelihood of false positives and negatives, and consequently improving patient trust and treatment outcomes.

Cost Savings: By identifying malignant tumors earlier, the treatment can be less invasive and more cost-effective, reducing the financial burden on patients and healthcare systems

The dataset used is the Wisconsin Breast Cancer Dataset, which contains 569 samples and 32 features.

Link to Dataset - Wisconsin Breast Cancer Dataset

The target variable - diagnosis - indicates whether the tumor is benign (B) or malignant (M)

data: Contains the dataset file.

notebooks: Jupyter Notebooks for data exploration, modeling, and evaluation.

scripts: Reusable Python code for model training and evaluation.

results: Exported images and results from model evaluations

This project was built using Python and the following key libraries:

pandas, numpy, matplotlib, seaborn, scikit-learn

Clone the repository.

Install required dependencies using pip install -r requirements.txt.

Run the Jupyter Notebook named notebooks/breast_cancer_detection.ipynb.

Using the Stratified cross-validation score, the SVC model outperformed the others with a score of 97.80%, making it the most reliable model for this dataset. This high accuracy suggests that the model is well-suited for initial breast cancer screening in a clinical setting.

Hyperparameter Tuning: Experiment with different model hyperparameters to improve predictive accuracy.

Feature Engineering: Remove or combine features with high self-correlation and low correlation with the target variable to reduce overfitting and improve model interpretability.

Model Deployment: Integrate the model into a web-based application for real-time predictions, enabling broader access for healthcare providers.

Contributions are welcome! Please follow these steps to contribute:

Fork the repository and create a new branch

Implement your changes and commit them with a clear message

Push to the branch and open a Pull Request for review.

This project is licensed under the MIT License. See the LICENSE file for more details.