Underfitted— Generalized — Overfitted

A brief note on how bias and variance makes a model as Underfitted or Generalized or Overfitted!

In this post, instead of writing so many paragraphs I just made an info-graphic for ease of understanding.

Image by Author

Underfitted:

A model could fit the training and testing data very poorly (high bias and low variance) — left most graph in above Info-graphic. This is known as underfitted.

Overfitted:

A modelcan fit the training data very well and the testing data very poorly. (low bias and high variance) — Right most graph in above Info-graphic. This is known as overfitted.

Generalized:

An ideal model should be low in Bias and also low in Variance.

Conclusion:

This post is a very small one because I do not want readers to confuse more.

Please leave your comments if you have any other idea on Bias-Variance, Overfitting & Underfitting.

If you want to dig more into Bias — Variance please read my previous blog Overfitting — Bias — Variance — Regularization.

Happy Programming!

Like to support? Just click the heart icon ❤️.

1+

Leave a Reply

Your email address will not be published. Required fields are marked *