Next · Generalisation and evaluationYou are hereReadBias and variance of a modelRefit one model on many training sets from one curve: a too-simple model misses the same way each time, a too-flexible one changes each time.