Prototypes within Minimum Enclosing Balls
We revisit the kernel minimum enclosing ball problem and show that it can be solved using simple recurrent neural networks. Once solved, the interior of a ball can be characterized in terms of a function of a set of support vectors and local minima of this function can be thought of as prototypes of the data at hand. For Gaussian kernels, these minima can be naturally found via a mean shift procedure and thus via another recurrent neurocomputing process. Practical results demonstrate that prototypes found this way are descriptive, meaningful, and interpretable.
- Published in:
ICANN 2019: Artificial Neural Networks and Machine Learning – ICANN 2019: Workshop and Special Sessions International Conference on Artificial Neural Networks (ICANN) - Type:
Inproceedings - Authors:
C. Bauckhage, R. Sifa, T. Dong - Year:
2019
Citation information
C. Bauckhage, R. Sifa, T. Dong: Prototypes within Minimum Enclosing Balls, International Conference on Artificial Neural Networks (ICANN), ICANN 2019: Artificial Neural Networks and Machine Learning – ICANN 2019: Workshop and Special Sessions, 2019, https://doi.org/10.1007/978-3-030-30493-5_36, Bauckhage.etal.2019a,
@Inproceedings{Bauckhage.etal.2019a,
author={C. Bauckhage, R. Sifa, T. Dong},
title={Prototypes within Minimum Enclosing Balls},
booktitle={International Conference on Artificial Neural Networks (ICANN)},
journal={ICANN 2019: Artificial Neural Networks and Machine Learning – ICANN 2019: Workshop and Special Sessions},
url={https://doi.org/10.1007/978-3-030-30493-5_36},
year={2019},
abstract={We revisit the kernel minimum enclosing ball problem and show that it can be solved using simple recurrent neural networks. Once solved, the interior of a ball can be characterized in terms of a function of a set of support vectors and local minima of this function can be thought of as prototypes of the data at hand. For Gaussian kernels, these minima can be naturally found via a mean shift...}}