Bias and variance tradeoff

In machine learning, we often use the least square difference to evaluate the model quality. Generally, the smaller the least square difference, the better the model. The least square difference is actually composed of bias and variance, and these two parts cannot be reduced at the same time. So in order to minimize the least square difference, we need to do some tradeoff between bias and variance.

Because of thisArticleIt involves many mathematical formulas, and csdn blogs do not support latex syntax. Therefore, they can only publish this article on their own blogs. For details, click bias and variance tradeoff.