Docstoc

The self-organizing map_ the Geo-SOM_ and relevant variants for

Document Sample
The self-organizing map_ the Geo-SOM_ and relevant variants for Powered By Docstoc
					 國立雲林科技大學
National Yunlin University of Science and Technology




 The self-organizing map, the Geo-SOM,
  and relevant variants for geosciences

                           Advisor : Dr. Hsu
                           Presenter : Wun-Cheng Cai
                           Author : Fernando Bacao, Victor Lobo
                                       Marco Painho



  2005,CG
                                                                                                 1
                                                              Intelligent Database Systems Lab
                                                               N.Y.U.S.T.
                                                                  I. M.

Outline

   Motivation
   Objective
   Method
   Experience
   Practical application
   Conclusion
   Personal Comments

     2
                            Intelligent Database Systems Lab
                                                                            N.Y.U.S.T.
                                                                               I. M.

Motivation

   We used SOM dealing with geo-referenced data, but
    do not consider geographical coordinate.




     3
                                         Intelligent Database Systems Lab
                                                                             N.Y.U.S.T.
                                                                                I. M.

Objective

   We present a new SOM architecture, the Geo-SOM,
    which was especially designed to take into account
    spatial dependency.




     4
                                          Intelligent Database Systems Lab
   X be the set of n training patterns
    x1,x2,..xn, each of these having a set
                                              Method                                 N.Y.U.S.T.
                                                                                        I. M.
    of components geoi and another set
    ngfi.
   W be a p x q grid of units wij where i
    and j are their coordinates on that
    grid, and each of these units having a
    set of components wgeoij and another
    set wngfij.
   αa be the learning rate, assuming
    values in [0,1], initialized to a given
    initial learning rate
   r be the radius of the neighborhood
    function h(wij,wmn,r), initialized to a
    given initial radius
   k be the radius of the geographical
    BMU that is to be searched
   f be a logical variable that is true if
    the units are at fixed geographical
    locations.




             5
                                                  Intelligent Database Systems Lab
                                                         N.Y.U.S.T.
                                                            I. M.

   Experience




Z=10            Z=0
       6
                      Intelligent Database Systems Lab
                                                          N.Y.U.S.T.
                                                             I. M.

Experience
             Geo-SOM                    SOM




   7
                       Intelligent Database Systems Lab
                                                           N.Y.U.S.T.
                                                              I. M.

Practical application


Geo-SOM      SOM




   8
                        Intelligent Database Systems Lab
                                                                             N.Y.U.S.T.
                                                                                I. M.

Conclusion

   Geo-SOM suited to deal with geo-referenced data.
   We present the Geo-SOM, which enable the treatment
    of the spatial dimension of the data separately from
    other attribute data, when performing data reduction
    tasks.




     9
                                          Intelligent Database Systems Lab
                                                                                      N.Y.U.S.T.
                                                                                         I. M.

Personal Comments

   Applications
    ─    understand geographical state of developing
   Advantages
    ─    deal with geo-referenced data
   Disadvantage
    ─    high quantization error




        10
                                                   Intelligent Database Systems Lab

				
DOCUMENT INFO
Shared By:
Categories:
Tags:
Stats:
views:3
posted:6/26/2012
language:
pages:10