Write a code named gd. m to implement gradient descent algorithm for the logistic regression problem (1) using the constant step-size scheme. the input for gd. m should include xtr, ytr, yte, yte, the regularization parameter λ, and the constant step-size t. the output for gd. m should be the training error, test error and the objective value at each iteration.
the code should appropriately address two critical issues:
i) how do you choose the constant step-size t?
ii) when do you think to stop the iteration process in your gd algorithm such that its iterative output is guaranteed to be sufficiently close to the optimal solution?