SequentialDownSampler

Sequential down sampling implementation for imbalanced data classification.

Usage:

Install Dependencies w pip:

1- numpy

Import the SequentialDownSampler Class and Create an Instance with Desired Settings.

from SequentialDownSampler import SequentialDownSampler
import numpy as np

# create a fake dataset.
X = np.random.rand(1000, 224, 224, 3)
Y = np.random.randint(0, 2, size=(1000))

# initiate the downsampler selecting the class to downsample and the ratio.
seqds = SequentialDownSampler(X, Y, dclass=1, ratio=3)

Perform Downsampling on your Dataset.

# perform downsampling.
X, Y = seqds.downsample()

print(X.shape)
print(Y.shape)