May I ask what does the 1/2 means in this formula for lasso. I searched formula examples for lasso and ridge. And found sometimes there is 1/2 while sometimes no.
2 Answers
It's there to cancel out the 2 from the power when you take the derivative. It's just for convenience thought, you can do it without it, but the math will look a bit uglier.
Typically, 1/2 is used in front of the MSE term or the square-regularisation terms, just to make the subsequent analysis more intuitive. The corresponding regularisation coefficient, $\lambda$, will be different but 1-1.