CrossValidation with fixed partitions

[haifengl]

CrossValidation with fixed partitions

[haifengl]

What do you mean "fixed partitions"?

[owlmsj]

"Running example" at 5-fold cross validation (partitions A,B,C,D and E = the entire dataset).

fold 1) A,B,C,D = train and E = test
fold 2) A,B,C,E = train and D = test
fold 3) A,B,D,E = train and C = test
fold 4) A,C,D,E = train and B = test
fold 5) B,C,D,E = train and A = test

Probably this is not the best explanation and it is kind of a trivial code to do. Just to know if this feature in smile is already coded or should I implement something look like and make a pull request...

[haifengl]

smile.validation.CrossValidation is for that.

[owlmsj]

My mistake (because my code was shuffling the data at the beginning). Sorry.