5,322
edits
No edit summary |
|||
Line 268: | Line 268: | ||
===Bias-Variance Tradeoff=== | ===Bias-Variance Tradeoff=== | ||
* Let <math>L_D(h)</math> be the true loss of hypothesis <math>h</math> and <math> | * Let <math>L_D(h)</math> be the true loss of hypothesis <math>h</math> and <math>L_s(h)</math> be the empirical loss of hypothesis <math>h</math>. | ||
** Here D is the true distribution and s is the training sample. | |||
* <math>L_D(h_s^*) = L_D(h_D^*) + [L_D(h_s^*) - L_D(h_D^*)]</math> | * <math>L_D(h_s^*) = L_D(h_D^*) + [L_D(h_s^*) - L_D(h_D^*)]</math> | ||
* The term <math>L_D(h_D^*)</math> is called the bias | * The term <math>L_D(h_D^*)</math> is called the bias |