Avoiding Overfitting mission: Variance measure


I’m a bit confused after reading the theory in this mission: Avoiding Overfitting:thinking:

When we treated the meassurements of Bias and Variance in the Cross Validation mission, we used the average of root mean square error (RMSE) values for meassuring bias, and standar desviation of these RMSE as variance meassurement. At this point ok, more or less clear, because the subject is complex… :sweat_smile:

However in this mission the variance of predicted values is used for meassuring the variance of the model instead. Sincerelly I can’t see the relation between this value, and the variance related to the total error of a model.

Here is the section at this mission where variance of predicted values is used: https://app.dataquest.io/m/132/overfitting/3/bias-variance-tradeoff

Anybody can explain me, or give me some link to more information about why they are using this different variance to calculate the variability of the prediction in this model?

Thanks in advance!


Hi @Daniel_H,

You can read this thread on StackExchange: Why underfitting is called high bias and overfitting is called high variance? Hope it helps!

1 Like

Many thanks! Great resource! :+1:t3:

1 Like