ON THE USE OF SELF ORGANIZING-MAPS FOR THE REPRESENTATION OF BARCODING DATA: AN APPLICATON TO HYLOMYSCUS DATA

Olteanu, M. (1), Violaine, N. (2), Schaeffer, B. (3), LAREDO, C. (3), David, O. (3)

(1) SAMM (EA4543), Université Paris 1 Panthéon Sorbonne, Paris, France
(2) Muséum National d’Histoire Naturelle, Paris, France
(3) INRA, France


Poster in Data Analysis Methods
Poster Location: D8


Self-organizing maps are used for visualizing Hylomyscus Barcoding data. The algorithm provides a two-dimensional projection of the samples and highlights the proximities and the dissimilarities between species, as well as the intra-specific variability. Since self-organizing maps are unsupervised learning, the available information on the species is only considered a posteriori. This way, it is possible to assess the hypothesis on the distribution of species and detect possible errors in the labeling of the samples.


Keywords: Vertebrates, Data Analysis