error来自bias 和variance
PARRELLEL UNIVERSE
假定每个平行宇宙里的我抓了十个不同的宝可梦
越复杂的model受data影响越大。
underfiting:not fitting the training examples
overfitting:large error on testing data
large bias:redesign the model:add more features as input/a more complex model
large variance:more data/regularization(可能伤害bias)
MODEL SELECTION
NOT do: