I've been reading recently about over-fitting and it is frequently related to High Sampling Variance and Low Bias characteristics.

However, what is the metric used to state the High Sampling Variance?

Take from example this polynomial function from the Over-fitting article of Wikipedia.


Is the Variance calculated by taking points randomly in the curve and taking its mean. Then, repeating this process X amount of times and calculate the Variance of the means?

Also, if possible, the answer has also to justify why the linear regression would have a smaller variance.

If it helps, I will link all the sources I've gather and read which will help us to solve this questioning.


Overfitting 1

Overfitting 2

Meaning Sampling Variance

Thank you in advance


As this video described https://www.youtube.com/watch?v=EuBBz3bI-aA

Variance would be calculated by taking a training set and the curve of prediction, and taking the sum of the squares between them.

| cite | improve this answer | |

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.