This repository corresponds to the code used for our journal submission to BMC Biomedical Informatics and Decision Making. You may have to cite the journal article as following if you were to use the data:
@article{yelugam2025formal,
title={A formal explanation space for the simultaneous clustering of neurologic diseases based on their signs and symptoms},
author={Yelugam, Raghu and Hier, Daniel B and Obafemi-Ajayi, Tayo and Carrithers, Michael D and Wunsch II, Donald C},
journal={BMC Medical Informatics and Decision Making},
year={2025},
publisher={Springer}
}
- Feel free to use the code, it's public for a reason!
- If you notice bugs, buy me a bugspray by reporting them.
- This code attempts to use various explanation methods applied two biclustering methods, which can be applied to any methods. Our data has labels, thus we used various CVIs, but if the data does not contain ground truth, so of the explanation may not relevant.
- Also, we used row labels for our analysis, this can be easily extended to column labels obtained upon biclustering.
- You may need to download biclustlib and use it if you wish to reproduce the comparison we generated. Note that biclustlib library is old and may not work well with latest versions of Python.
#1: Data provided here was sourced from medical text books, records, and other exemplars. Thus, no human subjects were involved. #2: The Research Square version may use both provided datasets, but the published version uses Dementia datasets per reviewer request.
Needs the following python libraries:
UV
use UV to install the requirements and create an environment.