/sci/ - Science & Math » Thread #11517366

53KiB, 720x904, 1585084425437.jpg

View Same Google iqdb SauceNAO

Anonymous Tue 31 Mar 06:25:19 2020 No.11517366 View Reply Original Report

Quoted By: >>11519238

>lower error rates on the validation set than on the training set
What did statistics mean by this?

Anonymous

Anonymous Tue 31 Mar 2020 06:34:51 No.11517377 Report

Quoted By: >>11518650

I mean, sometimes it does happen. Generally a sympton of small data sets though

Anonymous

Anonymous Tue 31 Mar 2020 18:13:45 No.11518650 Report

Quoted By: >>11520370

>>11517377
It's fitting 770k samples

Anonymous

Anonymous Tue 31 Mar 2020 18:18:42 No.11518660 Report

Quoted By:

Data leakage

Anonymous

Anonymous Tue 31 Mar 2020 21:40:55 No.11519238 Report

Quoted By:

>>11517366
You swapped your "training" and "validation" labels accidently.

Anonymous

Anonymous Wed 01 Apr 2020 04:39:30 No.11520370 Report

Quoted By:

>>11518650
Are you only fitting the model once or twice during validation? Fucking simp, 10 fold cross validation is the bare minimum for validation.

Capcode	All Only User Posts Only Moderator Posts Only Admin Posts Only Developer Posts
Show Posts	All Only With Images Only Without Images
Deleted Posts	All Only Deleted Posts Only Non-Deleted Posts
Ghost Posts	All Only Ghost Posts Only Non-Ghost Posts
Post Type	All Only Sticky Threads Only Opening Posts Only Reply Posts
Results	All Grouped By Threads
Order	Latest Posts First Oldest Posts First