Learning curves are useful in analyzing a machine learning model’s performance over various sample sizes of the training dataset.

To understand learning curves, it is important to have a good understanding of the Bias-Variance Tradeoff. You can check out my article regarding the same.

---

Evaluating Models

“Always plot learning curves while evaluating models”

Okay, so the basic thing we know is, if a model performs well on the training data but generalizes poorly, then the model is overfitting. If it performs poorly on both, then it is underfitting.

The hyperparameters must be set in such a way that, both bias and variance are as low as possible.

#learning-curve #machine-learning #data-science #artificial-intelligence #learning