the points that were not predicted correctly, i would think is associated with regions. When I have data looking like that is often the case that each region carries a different fitted line
This video is really appreciable! It would be interesting explaining how random forests/decision trees work. I was wondering also if it is possible to use seed to eliminate randomness in training or not, anyway keep going! 😊
You can definitely use a seed. Just set the random_state argument to a number. I also plan on making a video about decision trees and random forests from scratch in the future.
I've been practicing creating models lately, its been interesting and less frustrating surprisingly. Just a slight concern on your initial scores, mae and rmse should take the arguments in the following order (y_true, y_test), not sure if this would have a significant effect on the scores but it might just be.
It would be best practice to change the order, yes. But it does not affect the result in this case because the magnitude of the error is the same and the direction does not matter since we square the difference for the RMSE and we take the absolute value for the MAE. But nevertheless, you are right, the order should be swapped for consistency.
A small suggestion : why don't you try making videos on the Playground competitions in Kaggle as a series . Playground competitions because they do not take much of your time though.
Have thought about this. But I am not sure if I would do it in the form of a prepared project like this one or if I would do it as an unedited live coding session similar to my Codewars content.
Always enjoy your videos, Concise and to the point
Thanks! That's the goal :)
Will you continue your series of implementing data structures in python? Been loving the 4 videos so far!
Yes, it's always two videos of the series and two other videos. That's the upload schedule.
@NeuralNine Good to know that more is coming!
the points that were not predicted correctly, i would think is associated with regions. When I have data looking like that is often the case that each region carries a different fitted line
Thanks boss, you're the best! Where can I find tutorials related to geospatial datasets?
This video is really appreciable! It would be interesting explaining how random forests/decision trees work. I was wondering also if it is possible to use seed to eliminate randomness in training or not, anyway keep going! 😊
You can definitely use a seed. Just set the random_state argument to a number. I also plan on making a video about decision trees and random forests from scratch in the future.
I've been practicing creating models lately, its been interesting and less frustrating surprisingly. Just a slight concern on your initial scores, mae and rmse should take the arguments in the following order (y_true, y_test), not sure if this would have a significant effect on the scores but it might just be.
It would be best practice to change the order, yes. But it does not affect the result in this case because the magnitude of the error is the same and the direction does not matter since we square the difference for the RMSE and we take the absolute value for the MAE. But nevertheless, you are right, the order should be swapped for consistency.
oh nvm I think it was just on the initial mae score
Keyword being absolute
A small suggestion : why don't you try making videos on the Playground competitions in Kaggle as a series . Playground competitions because they do not take much of your time though.
Have thought about this. But I am not sure if I would do it in the form of a prepared project like this one or if I would do it as an unedited live coding session similar to my Codewars content.
Make a video on deployment of the models on server.
I already have videos for this :)
Great video bro , We want more python finance videos, KEEP IT UP 🔥
Thank you! :)
Nice video., thanks
Thank you!
Great video sir, please make some industry level projects for python and ml both
What would you be interested in?
@@NeuralNine something related to manufacturing and computer vision
Thx_