Categorical Encoding converts non-numerical categories into numerical values for ML algorithms. Techniques range from simple label encoding to more sophisticated one-hot, target, or embedding-based encodings. The choice of encoding can significantly impact model performance and interpretability.