biaswrappers 0.3.1

a Regression Wrapper that combats bias in data for machine learning models

Bias Wrappers

Wrappers for standard multioutput machine learning models that apply progressive calibration to training to produce better testing results, with a bias factor. Used mainly to combat bias on seemingly random/biased data. Default models are Linear Regression with Gradient Descent (for regression), however, you can input your own machine learning models with the model param.

Fixes

Fixed Array/List Contradiction in regression, removed classifier for code compatibility, and removed a few print statements.

Removed classifier because the formula used only benefits regression problems.

Instructions

1. Install the package with pip:

pip install biaswrappers

2. Python Quickstart:

# Import Classifier/Regressor
from biaswrappers import regressor
from biaswrappers.baseline_tests import test_regression

# Initialize classifier/regressor and...
# Specify a model class with a fit and predict method as a param.
my_regressor = regressor.BiasRegressor()

# Use the baseline_tests module for comparable results
test_regression(model=my_regressor) # No return values, just prints results