Warfit-learn

A machine learning toolkit for reproducible research in warfarin dose estimation.

Read the paper on arXiv for free or on Science Direct with your institutional access.

Contents

• Features
• Installation
• Usage
• Citing this work

Features

• Seamless loading, cleaning, and preprocessing of the IWPC warfarin dataset.
• Standardised implementations of scoring functions.
  • Percentage patients within 20% of therapeutic dose (PW20)
  • Mean absolute error (MAE)
  • R² coefficient
  • Hybrid scoring functions
  • Confidence intervals
• Multithreaded model evaluation using standardised resampling techniques.
  • Monte-carlo cross validation
  • Bootstrap resampling
• Full interoperability with NumPy, SciPy, Pandas, Scikit-learn, and MLxtend.

Supports Python 3.6+ on macOS, Linux, and Windows.

Installation

pip install warfit-learn

Usage

For a detailed tutorial, see the Getting Started document.

Seamless loading and preprocessing of IWPC dataset

from warfit_learn import datasets, preprocessing
raw_iwpc = datasets.load_iwpc()
data = preprocessing.prepare_iwpc(raw_iwpc)

Full scikit-learn interoperability

from sklearn.linear_model import LinearRegression
from sklearn.svm import LinearSVR
from warfit_learn.estimators import Estimator
my_models = [
    Estimator(LinearRegression(), 'LR'),
    Estimator(LinearSVR(loss='epsilon_insensitive'), 'SVR'),
]

Seamless, multithreaded research

from warfit_learn.evaluation import evaluate_estimators
results = evaluate_estimators(
    my_models,
    data,
    parallelism=0.5,
    resamples=10,
)

Citing this work

If you use warfit-learn in a scientific publication, please consider citing the following paper:

G. Truda and P. Marais, Evaluating warfarin dosing models on multiple datasets with a novel software framework and evolutionary optimisation, Journal of Biomedical Informatics (2020), doi: https://doi.org/10.1016/j.jbi.2020.103634

BibTeX entry:

@article{Truda2020warfit,
  title = "Evaluating warfarin dosing models on multiple datasets with a novel software framework and evolutionary optimisation",
  journal = "Journal of Biomedical Informatics",
  pages = "103634",
  year = "2020",
  issn = "1532-0464",
  doi = "https://doi.org/10.1016/j.jbi.2020.103634",
  url = "http://www.sciencedirect.com/science/article/pii/S1532046420302628",
  author = "Gianluca Truda and Patrick Marais",
  keywords = "Warfarin, Machine learning, Genetic programming, Python, Supervised learning, Anticoagulant, Pharmacogenetics, Software",
}

Copyright

Copyright (C) 2019 Gianluca Truda

This program is free software: you can redistribute it and/or modify it under the terms of the GNU General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version. This program is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details. You should have received a copy of the GNU General Public License along with this program. If not, see https://www.gnu.org/licenses/.