dbv

Decision Border Visualizer for Classification Algorithms

Project description

A python package for Machine Learning Engineers who want to visualize a classification algorithm's decision border.

Installation

Use pip to install dbv by typing or copying the following command.

pip install dbv

License

This package is licensed under GPL-3.0 License.

Example usage

If you are working with a classification algorithm and 2-dimensional input data (e.g. tabular data with two input columns) and a target column, you can use the plot_classifier function to visualize the decision border. The usage of this function inside a Jupyter notebook is recommended.

1. Load the data

import pandas as pd

df = pd.read_csv(<path to .csv file>)
df.head()

The dataframe might look like this:

2. Define X, y, split data

from sklearn.model_selection import train_test_split

X = df[["age", "interest"]].values
y = df["success"].values

X_train, X_test, y_train, y_test = train_test_split(X, y, random_state = 0, test_size = 0.25)

3. Define model

from sklearn.naive_bayes import GaussianNB

model = GaussianNB()
model.fit(X_train, y_train)

print(model.score(X_test, y_test))

4. Visualize Decision Border with plot_classifier

Set parameter proba = False to see a hard border.

from dbv import plot_classifier

# Plotting decision border for train data
plot_classifier(model, X_train, y_train, proba = False, xlabel = "Age", ylabel = "Interest")

Set parameter proba = True to see a soft border.

from dbv import plot_classifier

# Plotting decision border for train data
plot_classifier(model, X_train, y_train, proba = True, xlabel = "Age", ylabel = "Interest")

The result looks like this:

Source code & further information

The source code is maintained at https://github.com/sveneschlbeck/dbv
There are also further information concerning the GPL license model, contributing guidelines and more...