One-Hot Encoder + ML Error Diagnosis

Requirements First 30 minutes: Diagnose five scenarios comparing training and test error. For each scenario, describe what the pattern indicates (e.g., overfitting, underfitting, data leakage, distribution shift between train and test sets, or noisy labels). Be prepared to discuss techniques for reducing overfitting. Last 45 minutes: Implement a one-hot encoder from scratch. The encoder must learn a mapping from categories to indices and convert categorical values into…