Record details

Title
    Coenocline reconstruction using graph theory and Bayesian probability data generator
Author
    Čejchan, Petr
Conference
    United Kingdom -Visegrad Frontiers of Science Symposium (4. : 21.02.2007-23.02.2007 : Krakow, Polsko)
Language
    anglicky
Publication type
    abstrakt
Source title - monograph
    United Kingdom - Visegrad Four Frontiers of Science Symposium
Pages
    S. 1-2
Notes
    Rozsah: 2 s.
    Výzkumný záměr: CEZ:AV0Z30130516
Subject category
    coenocline
    gradient anaysis
    palaeoecology
Keyword
    Bayesian
    Coenocline
    Data
    Generator
    Graph
    Probability
    Reconstruction
    Theory
    Using
Abstract (in english)
   The goal is to design a novel way to reconstruct the 'species response function (SRFs) of unitary organisms on an unidimensional environmental/time gradient solely from abundance data on species communities (`indirect gradient analysis'. The word 'indirect' means that the gradient itself is latent, is not accessible, or even is presumed. The study will focus on the topology of the problem, thus disregarding the quantitative measuring of 'sample scores'. Formally, this task can be termed 'seriation of fossil communities (assemblages)'. Bayesian inference is used here to "squeeze' the population density distributions out of observed species' abundances. A large number of generated hypotheses serve to derive probability density functions of sample ranks, which are the result of the analysis.
Contributor
    AV ČR Brno, Geologický ústav
Contributor code
    AV ČR, GLÚ
Source format
    U
Import date
    8. 8. 2012