Premium
Usefulness of the SOM algorithm for estimation of species distribution and significance in comparing habitats
Author(s) -
Penczak T.
Publication year - 2011
Publication title -
journal of applied ichthyology
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.392
H-Index - 62
eISSN - 1439-0426
pISSN - 0175-8659
DOI - 10.1111/j.1439-0426.2011.01867.x
Subject(s) - biomass (ecology) , abundance (ecology) , biology , habitat , representation (politics) , plane (geometry) , set (abstract data type) , intensity (physics) , scale (ratio) , component (thermodynamics) , division (mathematics) , data set , ecology , biological system , statistics , mathematics , computer science , geometry , optics , cartography , physics , arithmetic , politics , political science , law , programming language , geography , thermodynamics
Summary The self‐organising map (SOM, ANN algorithm) and the component plane (entitled the SOMs visualization technique) have an hexagonal structure (i.e. division into neurons and clusters created with the same data set), which emphasises how the number or biomass of each species fluctuates and changes; this is represented as changes in a grey scale intensity from dark (maximum number or biomass) through medium (medium number or biomass) to light (absent). The importance of each species is easily visible on a component plane, especially by enriching the graphic representation with a vertical‐bar numerically quantifying greyness intensity and adding a numerical measure as the indicator value (IndVal) and its significance, or ranges of abundance or biomass recorded in samples.