neupy.datasets.make_digits
- neupy.datasets.make_digits(n_samples=100, noise_level=0.1, mode='flip')[source]
Returns discrete digits dataset.
Parameters: - n_samples : int
Number of samples. Defaults to 100.
- noise_level : float
Defines level of a discrete noise added to the images. Noise level defines probability for the pixel to be removed. Value should be in [0, 1) range. Defaults to 0.1.
- mode : {remove, flip}
This option allow to specify how additional noise will modify each image.
- flip - Per every randomly selected pixel function flips binary value. 1 -> 0 and 0 -> 1.
- remove - Per every randomly selected pixel function checks if value equal to 1 if it’s true that it gets replaced with 0.
Returns: - tuple
Tuple contains two values. First one is a matrix with shape (n_samples, 24). Second one is a vector that contains labels for each row. Each digit can be transformed into (6, 4) binary image.
Examples
>>> from neupy import datasets, utils >>> >>> utils.reproducible() >>> >>> digits, labels = datasets.make_digits(noise_level=0.15) >>> digit, label = digits[0], labels[0] >>> >>> label 5 >>> >>> digit.reshape((6, 4)) array([[0, 0, 1, 1], [1, 0, 0, 0], [0, 1, 1, 0], [0, 0, 0, 0], [0, 0, 0, 1], [1, 1, 1, 0]], dtype=uint8)