This exercise is way too complicated to implement especially on the school computers.
First of all there is a huge gap between the previous exercises and this one. Why isn't it more progressive?
Second of all we're supposed to test 64 different models with up to 17 different variables. On ex08 we gave us starting points for theta to speed up the process, why not do this for this exercise as well?
I guess the only way to make it work is to normalize the data but there is no exercise on this, not even a clue. We had a teaser on the last exercises of ML01 but we never learned how to use it, how to get back the real data, not even learned when to use standard or minmax normalization.
This exercise is way too complicated to implement especially on the school computers. First of all there is a huge gap between the previous exercises and this one. Why isn't it more progressive? Second of all we're supposed to test 64 different models with up to 17 different variables. On ex08 we gave us starting points for theta to speed up the process, why not do this for this exercise as well? I guess the only way to make it work is to normalize the data but there is no exercise on this, not even a clue. We had a teaser on the last exercises of ML01 but we never learned how to use it, how to get back the real data, not even learned when to use standard or minmax normalization.
Fixed on: