Generating noise in tabular data¶

This tutorial shows how to generate noise on pre-existing tabular data and to visualize both the original and the transformed data

In [1]:

from sklearn.datasets import make_blobs
from badgers.generators.tabular_data.noise import GaussianNoiseGenerator, GaussianNoiseClassesGenerator
import matplotlib.pyplot as plt
import numpy as np
import pandas as pd
import seaborn as sns

from sklearn.datasets import make_blobs from badgers.generators.tabular_data.noise import GaussianNoiseGenerator, GaussianNoiseClassesGenerator import matplotlib.pyplot as plt import numpy as np import pandas as pd import seaborn as sns

In [2]:

def plot_noise(X, y, Xt, yt):
    """
    Some utility function to generate the plots
    """
    fig, axes = plt.subplots(1, 2, sharex=True, sharey=True, figsize=(8,4))
    sns.scatterplot(data=X, x='dimension_0', y='dimension_1', hue=y, ax=axes[0])
    sns.scatterplot(data=Xt, x='dimension_0', y='dimension_1', hue=yt, ax=axes[1])
    axes[0].set_title('Original')
    axes[1].set_title('Transformed')
    return fig, axes

def plot_noise(X, y, Xt, yt): """ Some utility function to generate the plots """ fig, axes = plt.subplots(1, 2, sharex=True, sharey=True, figsize=(8,4)) sns.scatterplot(data=X, x='dimension_0', y='dimension_1', hue=y, ax=axes[0]) sns.scatterplot(data=Xt, x='dimension_0', y='dimension_1', hue=yt, ax=axes[1]) axes[0].set_title('Original') axes[1].set_title('Transformed') return fig, axes

Setup random generator¶

In [3]:

from numpy.random import default_rng
seed = 0
rng = default_rng(seed)

from numpy.random import default_rng seed = 0 rng = default_rng(seed)

Load and prepare data¶

We first load an existing dataset from sklearn.datasets

In [4]:

# load data
X, y = make_blobs(centers=4, random_state=0, cluster_std=0.25)
X = pd.DataFrame(data=X, columns=['dimension_0', 'dimension_1'])
y = pd.Series(y)

# load data X, y = make_blobs(centers=4, random_state=0, cluster_std=0.25) X = pd.DataFrame(data=X, columns=['dimension_0', 'dimension_1']) y = pd.Series(y)

Generate noise¶

The transformer applies an additive Gaussian noise to each dimension

In [5]:

trf = GaussianNoiseGenerator(random_generator=rng)
Xt, yt = trf.generate(X.copy(), y, noise_std=0.25)

trf = GaussianNoiseGenerator(random_generator=rng) Xt, yt = trf.generate(X.copy(), y, noise_std=0.25)

In [6]:

fig, axes = plot_noise(X, y, Xt, yt)

fig, axes = plot_noise(X, y, Xt, yt)

Generate noise for each class separately¶

The transformer applies an additive Gaussian noise to each dimension for each class separately.

In [7]:

trf = GaussianNoiseClassesGenerator(random_generator=rng)
Xt, yt = trf.generate(X.copy(), y, noise_std_per_class={0:0.1, 1:0.2, 2:0.3, 3:0.4})

trf = GaussianNoiseClassesGenerator(random_generator=rng) Xt, yt = trf.generate(X.copy(), y, noise_std_per_class={0:0.1, 1:0.2, 2:0.3, 3:0.4})

In [8]:

fig, axes = plot_noise(X, y, Xt, yt)

fig, axes = plot_noise(X, y, Xt, yt)

In [ ]: