Overfitting vs. Underfitting: Making Sense of the Bias-Variance Trade-Off | Towards Data Science

The best models live in the sweet spot: generalizing well, learning enough, but not too much