ex3 programming section - six random permutations of the training data?
How are we supposed to split the training data? choose 500 learning examples at random then test on the other 500? Should we try a different learning:test ratio for each permutation (i.e. 200:800, 900:100) or keep it fixed?
When running on X_test, should use the entire training sample for the learning stage?
Unfold
ex3 programming section - six random permutations of the training data? by anon (guest), 25 Nov 2013 00:44