Cross Validation

pykitml.cross_validate(inputs, outputs, folds=5)

Python generator function for making K-fold cross validation easier.

Parameters:

• inputs (numpy.array) – Inputs/features of training data.

• outputs (numpy.array) – Outputs/targets of training data.

Yields:

• train_inputs (numpy.array) – Training data containing inputs.

• train_outputs (numpy.array) – Training data containing outputs.

• test_inputs (numpy.array) – Testing data containing inputs.

• test_outputs (numpy.array) – Testing data containing outputs.

Example

>>> import numpy as np
>>> import pykitml as pk
>>>
>>> # Mock training data
... x = np.arange(30).reshape((10, 3))
>>> y = x + 10
>>>
>>> # 5-fold cross validation
... # Training data is split into 5 blocks, each block takes its turn
... # to be the test data.
... for train_x, train_y, test_x, test_y in pk.cross_validate(x, y, 5):
...     print(train_x)
...     print(train_y)
...     print(test_x)
...     print(test_y)