Problem with leave one out for neuralnet model

Olga92
Olga92 New Altair Community Member
edited November 2024 in Community Q&A
I'm working with the neural network model and I'm trying to validate using cross validation more precisely the leave one out mode, my problem is that I can't define the % I want to train and the rest to test. 
Initially I used a normal validation with a training percentage at the split ratio level.
How can I do it? 
Thanks for your help 
Tagged:

Answers

  • Telcontar120
    Telcontar120 New Altair Community Member
    When you use "leave one out" then you don't specify a train/test percentage because it will literally train on n-1 observations and test on a single observation at at time for n repetitions.  So if you have 100 examples, it will do this 100 times.  Be very careful with LOO cross valiation and large datasets...