- 4. (Loss Function) Generally speaking, a classifier can be written as H(x) = sign(F(x)), where H(x): Rd — -1,1 and F(x

[answerhappygod]

4 Loss Function Generally Speaking A Classifier Can Be Written As H X Sign F X Where H X Rd 1 1 And F X 1 (81.91 KiB) Viewed 13 times

- 4. (Loss Function) Generally speaking, a classifier can be written as H(x) = sign(F(x)), where H(x): Rd — -1,1 and F(x): Rd — R. To obtain the parameters in F(x), we need to minimize the loss function averaged over the training set: £ 1(y+F(x')). Here L is a function of yF(x). For example, Σ for linear classifiers, F(x) = Wo + £ -1 WjXj, and yF(x) = y(wo + Σ = wix) d (a) (20 points) Which loss functions below are appropriate to use in classification? For the ones that are not appropriate, explain why not. In general, what conditions does I have to satisfy in order to be an appropriate loss function? The x axis is yF(x), and the y axis is L(yF(x)). (a) (b) (c) sal al (a) (e) (6) (20 points) of the above loss functions appropriate to use in classification, which one is the most robust to outliers? (A model is robust when the parameters does not change significantly under various conditions.) Explain your answer.