logistic-bioem

Logistic Regression on Melatonin Biodata by Nicolas Bouché (nicolas.bouche@univ-lyon1.fr).

This is the algorithm used to make the figures used in the analysis in Bouché & McConway, Bioelectromagnetics 2019 (in press).

It uses PyMC3 and requires python2.7 or 3.5 (in a virtual environment). See INSTALL for instruction.

The figures of the paper can be made using the script below.

It is archived under https://zenodo.org/record/3250993 or https://doi.org/10.5281/zenodo.3250993

Run from logistic-bioem directory with

cd logistic-bioem;

ipython

from logistic import run_all
run_all.main(run=True,outpath='paper/')

Run as script from logistic-bioem directory:

to recreate the figures from the paper use

cd logistic-bioem;
python logistic/run_all.py paper/

In addition, it is possible to experiment with the outlier rejection as follows

python logistic/run_all.py path_to_save --format Robust_LR      #to use robust LR
python logistic/run_all.py path_to_save --format Robust_LR05    #to use robust LR with p_out=0.5
python logistic/run_all.py path_to_save --format default        #to turnoff robust_LR

To rerun the figures without running the Baeysian code:

python logistic/run_all.py paper/ --read

This should create the macro file and the 5 figures.

make paper

cd paper
pdflatex bfield_bioem; bibtex bfield_bioem; pdflatex bfield_bioem; pdflatex bfield_bioem

Bugs & issues

Please send bug report and/or issues with installation to nicolas.bouche@univ-lyon1.fr