Deciding which split ratio to use in hind sight of MAE and RMSE values

Anything related to deep learning, machine learning, data science, and artificial intelligence. Course topic requests can go here.
Post Reply
postingposted
Posts: 1
Joined: Sat Jul 10, 2021 2:40 am

Deciding which split ratio to use in hind sight of MAE and RMSE values

Post by postingposted »

Hi, i am doing a regression problem with various models on a small dataset of less than 100 rows with 7 features. The MAE and RMSE are the scorers in the GridSearchCV using LOOCV.
Turns out that the test scores are better than the train scores for split ratio 0.05 and 0.1, but the opposite way for split ratios 0.2 to 0.6.
How do i interpret this to decide which split ratio to use? In this case less than 10% is better or greater is better?
I plotted below graphs to visualize the problem i tried to describe above, appreciate some guidance..
Image
Post Reply

Return to “General Discussion”