Repository for research into diagnosing ASD (autism spectrum disorder) using machine learning. This research has been written into a paper, available in ElasticRemurs.pdf.
Uses many resources from the PyKale library: https://github.com/pykale/pykale.
Pipeline adapted from: https://github.com/pykale/pykale/tree/main/examples/multisite_neuroimg_adapt.
The code in this repository:
- applies various methods of classification to diagnose ASD
- applies the Remurs[1] method to the classification task on the ABIDE dataset
- proposes Elastic-Remurs method, an extension of Remurs with an additional smoothing penalty
Future areas of investigation:
- treat the decision function output as a certainty measure rather than just output a class
- use phenotypic data in the classification using Remurs and Elastic-Remurs
After cloning the repository, first install dependencies.
pip install -r requirements.txt
Everything will now be run from the code folder:
cd code
Before running the classification pipeline, you can configure the pipeline by editing config.py.
After configuration, run the pipeline with:
python main.py
To view results, use the view_results.py script. This script should also be run from the code directory.
python scripts/view_results.py results/{your_result_file}.csv
This basic view of the results will only show the alpha, beta, gamma values and the accuracy score. This will be updated.
[1] Song, X., & Lu, H. (2017). Multilinear Regression for Embedded Feature Selection with Application to fMRI Analysis. Proceedings of the AAAI Conference on Artificial Intelligence, 31(1). https://doi.org/10.1609/aaai.v31i1.10871