MultiClass SVDD

Multiclass classification using Support Vector Data Description (SVDD).

MultiClass_SVDD

Multiclass classification using Support Vector Data Description (SVDD).

Matlab code.

A multiclass classification based on SVDD is presented in this MATLAB script. The code is a first attempt to describe and adapt to Matlab the useful but complicated problem of multiclass SVDD. The code is nonetheless functional and suitable for theoretical testing, with the intention of making it perform well in strongly applied cases.

NClass_SVDD_exe.m is the example code of the multiclass SVDD, while NC_SVDD_TRAINING.m and NC_SVDD_TEST.m are the training and testing functions of the model respectively.

Other functions have been included to allow the program to function properly.

Cite As

-) Plain text

A. Carlevaro and M. Mongelli, "A New SVDD Approach to Reliable and Explainable AI," in IEEE Intelligent Systems, vol. 37, no. 2, pp. 55-68, 1 March-April 2022, doi: 10.1109/MIS.2021.3123669.

-) BibTex

@article{article,author = {Carlevaro, Alberto and Lenatti, Marta and Paglialonga, Alessia and Mongelli, Maurizio},year = {2023},month = {01},pages = {1-11},title = {Multi-Class Counterfactual Explanations using Support Vector Data Description},volume = {5},journal = {IEEE Transactions on Artificial Intelligence},doi = {10.1109/TAI.2023.3337053}}

ORCID: https://orcid.org/0000-0002-7206-5511