Need some assistance understanding a basic stochastic gradient descent (SGD)/Hinge Loss problem . In the problem below what are the steps I would go through to solve this?

4 restaurant reviews each labeled positive (+1) or negative (?1)

(?1) pretty smelly

(+1) good food

(?1) not good

(+1) pretty scenery

Each restaurant reviewxis mapped onto a feature vector?(x) which maps each word to the number of occurrences of that word in the review. For example, the first review maps to the (sparse) feature vector?(x)={pretty:1,smelly:1} . Recall the definition of the hinge loss:

\f\f\f