I recently discovered fastai and I’ve finished Part 1 of the course. Before moving on to Part 2 I’m having fun participating in Kaggle competitions and experimenting with the little I know.
So I’ve recently entered this competition:
I noticed there was a pinned notebook describing a method using random forest to get a baseline score.
They got a score of 0.633 which is better than all of my attempts.
Naturally, I tried to reproduce this result using sklearn’s random forests since it was in the course ( in the pinned notebook they used TensorFlow).
Using the code found in the fastbook as a base I managed to put together this simple notebook:
However this gets me a score of 1.194 which is not what I was expecting… I’m pretty sure I used the same method as the pinned notebook so I don’t understand why my score is so low.
Can someone help me understand why my random forest is not performing as theirs ?
Also I’m sorry if this is in the wrong category. I know this isn’t technically Deep Learning but I didn’t find a better category for my post.
Thank you for your help.