decorrelator 1.0.0

Easy-to-use Bayesian full and partial correlation estimator

Decorrelator

Analysing correlations between parameters whilst controlling for the effects of others.

(Fancy partial correlation)

Installation

• create new environment

  • conda create -n NAME python=3.7

• activate

  • source activate NAME

• install high-performance libraries

  • conda install numpy mkl scipy

• install correlator

  • pip install git+https://github.com/philastrophist/decorrelator.git

Use

Please see example.py for an example which uses nearly all functionality

There are two objects of interest: * decorrelator.LinearRelation * decorrelator.CorrelationModel

You may use them together as demonstrated in example.py or indivdually as shown in their home modules under if __name__ == "__main__"