When My First ML Model Memorized Instead of Learning (And How I Fixed It)

The author shares their experience of encountering overfitting in early ML projects, where models memorized training data instead of learning generalizable patterns. They resolved this by applying SMOTE for imbalanced data, K-Fold Cross Validation for reliable evaluation, and tuning model complexity in Random Forest. The key takeaway is that generalization on unseen data matters more than high training accuracy.