Train test split with every kind of cross-validation

This example shows how to split between test and train according to every cross-validation method.

Import librairies

import numpy as np
import museotoolbox as mtb

Generate random dataset

np.random.seed(42)
y = np.random.randint(1,3,10)
X = np.random.randint(1,255,[10,3],dtype=np.uint8)

Split train/test

Using museotoolbox.cross_validation.LeaveOneOut

cv = mtb.cross_validation.LeaveOneOut(random_state=42)

X_train, X_test, y_train, y_test = mtb.cross_validation.train_test_split(cv,X,y)

Split train/test with groups

Generate group

groups = np.array([1, 1, 2, 3, 4, 2, 1, 1, 2, 3],dtype=int)

Using museotoolbox.cross_validation.LeaveOneSubGroupOut

cv = mtb.cross_validation.LeaveOneSubGroupOut(random_state=42)

X_train, X_test, y_train, y_test, g_train, g_test = mtb.cross_validation.train_test_split(cv,X,y,groups=groups)