Table of Contents

• Installation
• Dependencies
• Tutorials
• Usage

Installation

To install this repo as Python lib just run the following command:

pip install git+https://github.com/VolodymyrVozniak/feature-visualizer

Dependencies

scikit-learn>=1.2.0
pandas>=1.5.2
plotly>=5.7.0

Tutorials

1. For binary problem check this tutorial
2. For regression problem check this tutorial
3. For multiclassification problem check this tutorial

Usage

There is 1 main class for feature visualization:

• Visualizer - to visualize features for binary, regression and multiclassification problems

Examples

• Binary problem

from sklearn.datasets import load_breast_cancer
from croatoan_visualizer.visualizer import Visualizer


X, y = load_breast_cancer(return_X_y=True, as_frame=True)
df = X.assign(target=y)

vis = Visualizer(
    df=df.reset_index(),
    id_column="index",
    target_column="target"
)

vis.pca2d()
vis.pca3d()

vis.tsne2d()
vis.tsne3d()

For more details check tutorial

(back to top)

• Regression problem

from sklearn.datasets import load_diabetes
from croatoan_visualizer.visualizer import Visualizer


X, y = load_diabetes(return_X_y=True, as_frame=True)
df = X.assign(target=y)
df["target"] = pd.qcut(df["target"], 10, labels=False, duplicates='drop')

vis = Visualizer(
    df=df.reset_index(),
    id_column="index",
    target_column="target"
)

vis.pca2d()
vis.pca3d()

vis.tsne2d()
vis.tsne3d()

For more details check tutorial

(back to top)

• Multiclassification problem

from sklearn.datasets import load_iris
from croatoan_visualizer.visualizer import Visualizer


X, y = load_iris(return_X_y=True, as_frame=True)
df = X.assign(target=y)

vis = Visualizer(
    df=df.reset_index(),
    id_column="index",
    target_column="target"
)

vis.pca2d()
vis.pca3d()

vis.tsne2d()
vis.tsne3d()

For more details check tutorial

(back to top)