somplot: Visualisation of hexagonal Kohonen maps

Example dataset

The following dataset can be used to test the webservice and the package. It incudes a dataset of analytical data for 416 olive oil samples from nine different regions of Italy; for each sample the normalized concentrations of eight fatty acids are given.
The dataset is provided by Prof. Michele Forina, University of Genova and published by Gasteiger and Zupan in the book: J. Zupan, J. Gasteiger Neural Networks in Chemistry and Drug Design: An Introduction 2nd Edition, Wiley-VCH, Weinheim, 1999

The dataset is modified by limiting the number of oils for each region to a maximum of 50. For the full dataset and further information see the website of the book: Olive oil dataset in Johnny´s ANN-book.

Our demo archive includes the files oils-50.dat and oils-50.vis. The file oils-50.dat holds the data in SOM_PAK input format. somplot uses this file to extract the class information (annotation) for each record, which is stored as plain text in the last column.

The dataset has been trained with SOM_PAK tools and mapped to a self-organising map of hexagonal topology and a geometry of 15x45 codebook vectors. The file oils-50.vis. is the raw output of the SOM_PAK tool visual ran on the trained som. The R-script will read the mapping of oils to neurons from this file.

Download olive oils demo dataset: olive-oils.zip