python- Best techniques to split datase to get high performance accuracy

I apply this 4 method :

1. Train and Test Sets.
2. K-fold Cross Validation.
3. Leave One Out Cross
4. Validation. Repeated Random Test-Train Splits.

The method Train and Test Sets achieve high accuracy but the rest method achieve same accuracy but lower then first approah.

I want to know what is method i should choice?