Particionamiento con ShuffleSplit

• Ultima modificación: 2023-02-27 | `YouTube <>`__

#
# Parte aleatoriamente la muestra, sin garantizar que los grupos
# sean diferentes
#
from sklearn.model_selection import ShuffleSplit

shuffleSplit = ShuffleSplit(
    # --------------------------------------------------------------------------
    # Número de grupos
    n_splits=10,
    # --------------------------------------------------------------------------
    # Tamaño del conjunto de prueba
    # int: número de ejemplos
    # float: porcentaje de la muestra
    test_size=0.25,
    # --------------------------------------------------------------------------
    # Tamaño del conjunto de entrenamiento
    # int: número de ejemplos
    # float: porcentaje de la muestra
    train_size=None,
    # --------------------------------------------------------------------------
    # Semilla del generador de aleatorios
    random_state=0,
)

shuffleSplit

ShuffleSplit(n_splits=10, random_state=0, test_size=0.25, train_size=None)

from mymodule import plot_schema

y_classes = [0] * 10 + [1] * 10

plot_schema(shuffleSplit, y_classes)

import numpy as np

X = np.array([[1, 2], [3, 4], [5, 6], [7, 8], [9, 10], [11, 12]])
y = np.array([1, 2, 3, 4, 5, 6])

shuffleSplit = ShuffleSplit(
    n_splits=10,
    test_size=2,
    random_state=12345,
)

for i, (train_index, test_index) in enumerate(shuffleSplit.split(X)):
    print(f"Fold {i}:")
    print(f"  Train: index={train_index}")
    print(f"  Test:  index={test_index}")
    print()

Fold 0:
  Train: index=[4 0 1 2]
  Test:  index=[5 3]

Fold 1:
  Train: index=[3 1 2 5]
  Test:  index=[4 0]

Fold 2:
  Train: index=[5 2 3 1]
  Test:  index=[0 4]

Fold 3:
  Train: index=[3 1 2 0]
  Test:  index=[4 5]

Fold 4:
  Train: index=[5 4 2 1]
  Test:  index=[3 0]

Fold 5:
  Train: index=[0 3 1 5]
  Test:  index=[4 2]

Fold 6:
  Train: index=[1 0 4 3]
  Test:  index=[5 2]

Fold 7:
  Train: index=[4 1 5 3]
  Test:  index=[2 0]

Fold 8:
  Train: index=[1 0 3 5]
  Test:  index=[2 4]

Fold 9:
  Train: index=[0 5 3 4]
  Test:  index=[1 2]