I am working on through lesson 4 of Machine Learning: Zero to GBMs and REALLY enjoying the Random Forest and all of its parameters and how messing with them affects the model. The effects are not AS exciting as I would sometimes like them to be, but I have come up with a good little system of having a function run through a series of parameter settings and then plot out the results so that I can find the best possible parameter settings for the model.
It is SO MUCH FUN!
I thought I would share my work here in case anyone else finds it interesting: https://jovian.ai/evanmarie/sklearn-decision-trees-random-forests
I have not written a function to do the plotting yet, but I am starting to think that it would be REALLY useful to be able to just hit “Run” and the program go through all the settings and automatically plot so I can see not only what it considers to be the best outcome based on parameter settings but also so I can infer from the data surrounding the best outcome in case there is something the program might not have noticed. (UPDATE: I wrote the function. So cool! Now, I need to write one for plotting more than one parameter being tested at a time. Seeing the correlations of parameters is super interesting!)
Anyway, I could go on about this all day, because I love it so much!