Backwards Regression Python Library - Automated feature selection in linear and logistic regression models.

The Backwards Regression Python Library is an open-source toolkit for automated feature selection in regression models. It supports both linear and logistic regression, dynamically selecting the appropriate method based on the target variable.

Installation

pip install backwards_regression

Load Package

## Load Package
from backwards_regression import fit_logistic
from backwards_regression import fit_linear

Usage

## (Linear) With interactions included - set to True and Without Interactions included - set to False
result, dropped_vars = fit_linear(X, y, threshold_in=0.01, threshold_out=0.05, include_interactions=True, verbose=True)

## Print Selected features
print("Final included features:", result)

## Print Eliminated features
print("Dropped variables:", dropped_vars)

## (Logistic) With interactions included - set to True and Without Interactions included - set to False
result, dropped_vars = fit_logistic(X, y, threshold_in=0.01, threshold_out=0.05, include_interactions=True, verbose=True)

## Print Selected features
print("Final included features:", result)

## Print Eliminated features
print("Dropped variables:", dropped_vars)

Key Features

• Automated backward regression for linear and logistic regression models.
• Inclusion and exclusion of features based on user-defined significance thresholds.
• Optional inclusion of interaction terms for enhanced model complexity.

This library is suitable for data scientists, researchers, and practitioners working with regression problems who seek a streamlined approach to feature selection. The library intelligently adapts to the nature of the target variable, supporting both linear and logistic regression models.

Documentation & Examples

For documentation and usage examples, visit the GitHub repository: https://github.com/knowusuboaky/backwardsreg

Author: Kwadwo Daddy Nyame Owusu - Boakye
Email: kwadwo.owusuboakye@outlook.com
License: MIT