Bias-Variance trade-off in Ensemble Learning

When we discuss ensemble learners, one of the most important topics to discuss is the Bias-Variance trade-off. So, let’s dig in deeper into what actually that means.

We are going to analyse the following image because it offers a nice visualisation of what biais and variance mean.

Ze have on the left top Low Variance-Low Bias. That means that the points are centred on the bullseye. This will be the work of a pro dart player.

On the Low Bias-High Variance that means that you are centered arroud the bully, the dart player is not consistently hitting it. So Low Bias means that the points are centered arrouund the bullseye while High Variance means that the points are spread out and not consistently hitting whatever you’re aiming for, the bullseye in this case. So that means that the player is a fairly good dart player, but not quite good enough to consistently hit the bullseye.

Now High Bias-Low Variance this means you’re not centered around the bullseye, but the darts are fairly concentrated. This would either be somebody that’s confused about what they should be aiming for, or maybe it’s a pro dart player that’s just aiming for triple 20 or something other than the bullseye.

High Bias-High Variance the points are not centered around the bullseye and also they are…