Week 1 video bias / variance

[kpommois]

Hi!

I understood more or less the concept of high/low bias / variance but I did not get the following slide:

(one of the first slide of the video bias/ variance, when he plot the datasata and define high variace/bias and just right fitting)

First, I don’t understand what is plotting. Then, I don’t understand how he can define high variance and bias or if it is just right from those plot.

Can somebody help with this or suggest some refences to help me to understand this?

Thank you!

[kampamocha]

Hi @kpommois,

I understand it like this, the plots are about a hypothetical model with two features represented as x and y axes and two classes represented as circles and crosses. The drawed blue line represents the model that is trying to separate those classes.

At the left you have a model that is very simple and the expected value of its predictions is not very close to reality (high bias). At the right you have a model that is very complex and adjust very well to the existing data, but has much variability in its predictions so does not generalize well (high variance).

If you want to know more about this, @nramon has posted a very good explanation here:

Overfitting and underfitting at the same time ? Bias-variance-trade-off

Ideally, high variance models are complex and represent all the features of the training set very well leading to minimal error on the training set but fail to generalize to the unseen data. In contrast, high bias models represent extremely simple mappings and can generalize some features to the unseen data, but the simplicity of these models leads to underfitting on the training set and generates predictions with high bias when applied to data outside of the training set .But , If we took the q…