R exercise

Leave-One-Out Cross-Validation 2

Sign in to hand in your solution and get feedback.

cv.error <- ... for (i in 1:5) { glm.fit = ... cv.error[i] = ... }