Out-of-bag (OOB) error derivation for Random Forests

Рет қаралды 13,575

5 жыл бұрын

The RandomForestClassifier is trained using bootstrap aggregation, where each new tree is fit from a bootstrap sample of the training observations . The out-of-bag (OOB) error is the average error for each calculated using predictions from the trees that do not contain in their respective bootstrap sample. This allows the RandomForestClassifier to be fit and validated whilst being trained.
In this video, I example how you end up training only 64% of your training data when OOB option is set true in Random Forest.
If you do have any questions with what we covered in this video then feel free to ask in the comment section below & I'll do my best to answer those.
If you enjoy these tutorials & would like to support them then the easiest way is to simply like the video & give it a thumbs up & also it's a huge help to share these videos with anyone who you think would find them useful.
Please consider clicking the SUBSCRIBE button to be notified for future videos & thank you all for watching.
You can find me on:
GitHub - github.com/bhattbhavesh91
Medium - / bhattbhavesh91
#oob_error #randomforest #validationset #python #deeplearning #datascience #youtube

Пікірлер: 26

@fealgu100 3 жыл бұрын

Straight to the point. Your explanation gives me legs, now I run. Thanks.

@bhattbhavesh91 3 жыл бұрын

Thank you so much 🙂

@sushilchauhan2586 4 жыл бұрын

i am his huge fan coz every tym he covers that topic which is important but rarely available on youtube

@bhattbhavesh91 4 жыл бұрын

Thanks!

@meanuj1 5 жыл бұрын

Thanks very much for video...pls keep making ..

@whenmathsmeetcoding1836 4 жыл бұрын

Thanks you so much awesome content. keep it up...

@MrBungwung 5 жыл бұрын

I understand your math, but why are you assuming that "n" for selecting the row is equal to the "n" rows you chose from?

@xuanzhiong3757 4 жыл бұрын

I am having the same question.

@VIGNESHPRAJAPATI 5 жыл бұрын

Nice video.

@amandhimanchroma 5 жыл бұрын

Beautifully proved.

@meswapnilspal 4 жыл бұрын

in expresion (1 - 1/n)^n , how come both variables are same, why is the expression is not, (1 - 1/a)^b, where a -> no. of row in dataset, and b -> no. of times we choose to run the decision tree.?

@shambhaviaggarwal9977 4 жыл бұрын

even I have the same question.

@user-ko8or1lr7v 4 жыл бұрын

@@shambhaviaggarwal9977 Couse we consider only one subset, Out-Of-Bag error counts for EVERY subset. So, for ONE subset prob = (1 - 1/(# rows))^(# rows), if length of subset = length of original set

@ahmadsaufie3542 3 жыл бұрын

Nice video... Can you explain, how OOB can be use to determine the optimal number of iteration for RF?

@NotFound-hy7qb 5 жыл бұрын

Cant we use cross validation instead of oob ?

@bhavikdudhrejiya4478 4 жыл бұрын

Very nice way to explained Do you have any links for this?

@nailashah6918 4 жыл бұрын

Can we find the weight of a tree based on out of bag error?

@tototong1 4 жыл бұрын

lim(y->0) 1/y*ln(1+y) ===> d/dy(N^r) = d/dy ln(1+y) Could you explain more about this? Thanks!!!

@meanuj1 5 жыл бұрын

One doubt this oob is attribute which we can set in random forest?? If yes then how to get that oob data for validation ??

@bhattbhavesh91 5 жыл бұрын

Hi Anuj! Thank yo so much for your kind words! I have made a video on how you can code this up, I'll be publishing the video on next Thursday! So, stay tuned :)