Consider linear regression y = Xw where y, X and w have size n 1, n p and p 1. you wanted the components of w to be nonnegative. Explain how you might do this. You can provide R code that uses optim if you think that is helpful, but this is not necessary