Introduction

This code can be used to replicate the results reported in the paper: "Modeling the Impact of Inter-Rater Disagreement on Sleep Statistics using Deep Generative Learning" (Accepted for publication in J-BHI).

Environment

To make use of the environment, first install Anaconda. In anaconda prompt navigate to the parent directory and run:

conda env create -f environment.yml --prefix ./env

and then activate the environment:

conda activate ./env

unzipping the predictions

The run the comparison and plotting scripts, please first unzip "predictions.zip" in the root directory

Compare

To recreate the results as listed in the tables of the manuscript, run:

python Compare.py

Optionally one can specify which method to compare by passing their names as an argument:

python Compare.py -names_to_compare= U-Net_fact U-Flow 

plot

To recreate the plots of the manuscript, run:

python plot.py

Optionally one can specify which method to compare by passing their names as an argument, additonally the subject to plot can also be specified:

python plot.py -names_to_compare U-Net_fact U-Flow  -subject_id 19

preprocess

To mimic the preprocessing performed on the IS-RC dataset, first download the dataset from here: link.

Then, prepocessing the dataset can be performed by running:

python preprocess.py -data_loc="location//of//the//data"

with the data_loc argument pointing towards the location of the raw IS-RC dataset.