biaswrappers 0.3

Wrappers that combat 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) and a standard Naive Bayes (for classification), however, you can input your own machine learning models with the model param.

Fixes

Added a get_params function and switched default ml algs to sklearn framework to help with sklearn compatibility issues; therefore removing _models.py.

Instructions

1. Install the package with pip:

pip install biaswrappers

2. Python Quickstart:

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

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

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