
Project Overview
Used the Sonar dataset (60 features) to classify materials as rock or mine. Performed scaling, PCA-exploration and trained logistic regression and tree-based models with cross-validation.
Tools & Techniques
PythonPandasPCAScikit-learn
Key Outcomes
- 83% classification accuracy with regularized logistic regression.
- Feature-scaling and PCA improved model stability.
- Clear decision boundaries visualized for analysis.