search

mattlichti / gen_assembly

Repository for data science 10 course
0 stars 0 forks source link

HW3 Review #3

Open cdp13 opened 10 years ago

cdp13 commented 10 years ago

Hi Matt,

Your code is clean and easy to follow. Certainly seems to achieve the goal of visualizing the data and iterating over various factor combinations to discover the optimal predictive formula.

I am curious, though, how you decided to incorporate the squared term on problems completed. I am sure there is some logical reasoning behind it, but it would have been helpful to have that reasoning spelled out in comments.

And I loved your summary at the end! That was very helpful -- and practical.

Best, Caroline

PS: I had totally forgotten about instructor tags, @cdp13 @kebaler @ghego @craigsakuma

mattlichti commented 10 years ago

Hi Caroline,

I was planning on experimenting with a bunch of polynomial features in my model, but I didn't get around to trying anything other than the problems completed squared term. I didn't include it in my final model since it seemed kind of random to just have that one polynomial term, and it complicated things without improving fit much. I should have deleted that section or explained it more in the comments.

Thanks, Matt .