5.1.6. Exercises
- Regularization
Explain why the bias parameter \(\theta_0\) is not considered in the regularization cost.
What will be the result of linear regression in case \(\lambda\rightarrow\infty\) is chosen?
- Regularization
Show that
\[\pfrac{}{\v\theta} \left( \frac{\lambda}{2m} \|D_0\v\theta\|^2 \right) = \frac{\lambda}{m}D_0 \v\theta\]You can either calculate the gradient vector elementwise or use results from the section Matrix Calculus.