A Nonparametric Classification Algorithm based on Optimized Templates

0502361 - ÚI 2020 RIV CH eng C - Konferenční příspěvek (zahraniční konf.)
Kalina, Jan
A Nonparametric Classification Algorithm based on Optimized Templates.
Nonparametric Statistics : 3rd ISNPS, Avignon, France, June 2016. Cham: Springer, 2018 - (Bertail, P.; Blanke, D.; Cornillon, P.; Matzner-Lober, E.), s. 119-132. Springer Proceedings in Mathematics & Statistics, 250. ISBN 978-3-319-96941-1. ISSN 2194-1009.
[ISNPS 2016. Conference of the International Society for Non-Parametric Statistics /3./. Avignon (FR), 11.06.2016-16.06.2016]
Grant CEP: GA MZd(CZ) NV15-29835A
Institucionální podpora: RVO:67985807
Klíčová slova: supervised learning * nonparametric classification * nonlinear optimization * image analysis * object localization
Obor OECD: Computer sciences, information science, bioinformathics (hardware development to be 2.2, social aspect to be 5.8)

In this paper, a novel nonparametric classification method to two groups is proposed, which is based on a centroid (prototype, template) of one of the groups. The method does not consider any distributional assumptions, allows a clear interpretation and optimizes the centroid without any parametric model, as it is common in the nonparametric regression context.
Trvalý link: http://hdl.handle.net/11104/0294301