In k-fold cross validation, how do you estimate the test error in a regression problem? As each fold will end up with the same value of MSE, you can use this number as the estimated test error. Find the average MSEs that were measure k times on k sets of validation data (size = n/k). Find the average MSEs that were measure k times on the data used to train those k models (size = n - n/k) Find the sum of all MSE measurements on k sets of validation data