XSVMC-Lib is an open source library that implements algorithms proposed to make support vector machine (SVM) predictions interpretable.
XSVMC-Lib requires Python 3.8+. Since XSVMC-Lib has been implemented as an extension of sklearn.svm.SVC, it also requires the SciKit-Learn package (python3 -m pip install sklearn).
Although they are not required by XSVMC-Lib, the following packages are necessary for running the examples:
- Numpy (
python3 -m pip install numpy
) - Matploplib (
python3 -m pip install matploplib
) - OpenCV (
python3 -m pip install opencv-python
)
To run an example, say digits_explanation.py, you may use the following commands:
cd /path/to/xsvmc-lib
python3 examples/digits_explanation.py
Parts of the following datasets are used in several examples that illustrate the use of XSVMC-Lib.
The mathematical foundation of XSVMC-Lib can be found in
M. Loor and G. De Tré, "Contextualizing Support Vector Machine Predictions," International Journal of Computational Intelligence Systems, Volume 13, Issue 1, 2020, Pages 1483 - 1497, doi: 10.2991/ijcis.d.200910.002.
XSVMC-Lib is released under the Apache License, Version 2.0.
If you use XSVMC-Lib, please cite the following article:
M. Loor, A. Tapia-Rosero and G. De Tré, "An Open-Source Software Library for Explainable Support Vector Machine Classification," 2022 IEEE International Conference on Fuzzy Systems (FUZZ-IEEE), 2022, pp. 1-7, doi: 10.1109/FUZZ-IEEE55066.2022.9882731.
@INPROCEEDINGS{
xsvmlib,
author={Loor, Marcelo and Tapia-Rosero, Ana and {De Tr\'{e}}, Guy},
booktitle={2022 IEEE International Conference on Fuzzy Systems (FUZZ-IEEE)},
title={An Open-Source Software Library for Explainable Support Vector Machine Classification},
year={2022}, pages={1-7},
doi={10.1109/FUZZ-IEEE55066.2022.9882731}
}