Mercurial > hg > camir-aes2014

annotate toolboxes/MIRtoolbox1.3.2/somtoolbox/som_neighborhood.m @ 0:e9a9cd732c1e tip

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
first hg version after svn
author wolffd
date Tue, 10 Feb 2015 15:05:51 +0000
parents
children
rev   line source
wolffd@0 1 function Ne = som_neighborhood(Ne1,n)
wolffd@0 2
wolffd@0 3 %SOM_NEIGHBORHOOD Calculate neighborhood matrix.
wolffd@0 4 %
wolffd@0 5 % Ne = som_neighborhood(Ne1,n)
wolffd@0 6 %
wolffd@0 7 % Ne = som_neighborhood(Ne1);
wolffd@0 8 % Ne = som_neighborhood(som_unit_neighs(topol),2);
wolffd@0 9 %
wolffd@0 10 % Input and output arguments ([]'s are optional):
wolffd@0 11 % Ne1 (matrix, size [munits m]) a sparse matrix indicating
wolffd@0 12 % the units in 1-neighborhood for each map unit
wolffd@0 13 % [n] (scalar) maximum neighborhood which is calculated, default=Inf
wolffd@0 14 %
wolffd@0 15 % Ne (matrix, size [munits munits]) neighborhood matrix,
wolffd@0 16 % each row (and column) contains neighborhood
wolffd@0 17 % values from the specific map unit to all other
wolffd@0 18 % map units, or Inf if the value is unknown.
wolffd@0 19 %
wolffd@0 20 % For more help, try 'type som_neighborhood' or check out online documentation.
wolffd@0 21 % See also SOM_UNIT_NEIGHS, SOM_UNIT_DISTS, SOM_UNIT_COORDS, SOM_CONNECTION.
wolffd@0 22
wolffd@0 23 %%%%%%%%%%%%% DETAILED DESCRIPTION %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
wolffd@0 24 %
wolffd@0 25 % som_neighborhood
wolffd@0 26 %
wolffd@0 27 % PURPOSE
wolffd@0 28 %
wolffd@0 29 % Calculate to which neighborhood each map unit belongs to relative to
wolffd@0 30 % each other map unit, given the units in 1-neighborhood of each unit.
wolffd@0 31 %
wolffd@0 32 % SYNTAX
wolffd@0 33 %
wolffd@0 34 % Ne = som_neighborhood(Ne1);
wolffd@0 35 % Ne = som_neighborhood(Ne1,n);
wolffd@0 36 %
wolffd@0 37 % DESCRIPTION
wolffd@0 38 %
wolffd@0 39 % For each map unit, finds the minimum neighborhood to which it belongs
wolffd@0 40 % to relative to each other map unit. Or, equivalently, for each map
wolffd@0 41 % unit, finds which units form its k-neighborhood, where k goes from
wolffd@0 42 % 0 to n.
wolffd@0 43 %
wolffd@0 44 % The neighborhood is calculated iteratively using the reflexivity of
wolffd@0 45 % neighborhood.
wolffd@0 46 % let N1i be the 1-neighborhood set a unit i
wolffd@0 47 % and let N11i be the set of units in the 1-neighborhood of any unit j in N1i
wolffd@0 48 % then N2i (the 2-neighborhood set of unit i) is N11i \ N1i
wolffd@0 49 %
wolffd@0 50 % Consider, for example, the case of a 5x5 map. The neighborhood in case of
wolffd@0 51 % 'rect' and 'hexa' lattices (and 'sheet' shape) for the unit at the
wolffd@0 52 % center of the map are depicted below:
wolffd@0 53 %
wolffd@0 54 % 'rect' lattice 'hexa' lattice
wolffd@0 55 % -------------- --------------
wolffd@0 56 % 4 3 2 3 4 3 2 2 2 3
wolffd@0 57 % 3 2 1 2 3 2 1 1 2 3
wolffd@0 58 % 2 1 0 1 2 2 1 0 1 2
wolffd@0 59 % 3 2 1 2 3 2 1 1 2 3
wolffd@0 60 % 4 3 2 3 4 3 2 2 2 3
wolffd@0 61 %
wolffd@0 62 % Because the iterative procedure is rather slow, the neighborhoods
wolffd@0 63 % are calculated upto given maximal value. The uncalculated values
wolffd@0 64 % in the returned matrix are Inf:s.
wolffd@0 65 %
wolffd@0 66 % REQUIRED INPUT ARGUMENTS
wolffd@0 67 %
wolffd@0 68 % Ne1 (matrix) Each row contains 1, if the corresponding unit is adjacent
wolffd@0 69 % for that map unit, 0 otherwise. This can be calculated
wolffd@0 70 % using SOM_UNIT_NEIGHS. The matrix can be sparse.
wolffd@0 71 % Size munits x munits.
wolffd@0 72 %
wolffd@0 73 % OPTIONAL INPUT ARGUMENTS
wolffd@0 74 %
wolffd@0 75 % n (scalar) Maximal neighborhood value which is calculated,
wolffd@0 76 % Inf by default (all neighborhoods).
wolffd@0 77 %
wolffd@0 78 % OUTPUT ARGUMENTS
wolffd@0 79 %
wolffd@0 80 % Ne (matrix) neighborhood values for each map unit, size is
wolffd@0 81 % [munits, munits]. The matrix contains the minimum
wolffd@0 82 % neighborhood of unit i, to which unit j belongs,
wolffd@0 83 % or Inf, if the neighborhood was bigger than n.
wolffd@0 84 %
wolffd@0 85 % EXAMPLES
wolffd@0 86 %
wolffd@0 87 % Ne = som_neighborhood(Ne1,1); % upto 1-neighborhood
wolffd@0 88 % Ne = som_neighborhood(Ne1,Inf); % all neighborhoods
wolffd@0 89 % Ne = som_neighborhood(som_unit_neighs(topol),4);
wolffd@0 90 %
wolffd@0 91 % SEE ALSO
wolffd@0 92 %
wolffd@0 93 % som_unit_neighs Calculate units in 1-neighborhood for each map unit.
wolffd@0 94 % som_unit_coords Calculate grid coordinates.
wolffd@0 95 % som_unit_dists Calculate interunit distances.
wolffd@0 96 % som_connection Connection matrix.
wolffd@0 97
wolffd@0 98 % Copyright (c) 1999-2000 by the SOM toolbox programming team.
wolffd@0 99 % http://www.cis.hut.fi/projects/somtoolbox/
wolffd@0 100
wolffd@0 101 % Version 1.0beta juuso 141097
wolffd@0 102 % Version 2.0beta juuso 101199
wolffd@0 103
wolffd@0 104 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
wolffd@0 105 %% Check arguments
wolffd@0 106
wolffd@0 107 error(nargchk(1, 2, nargin));
wolffd@0 108
wolffd@0 109 if nargin<2, n=Inf; end
wolffd@0 110
wolffd@0 111 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
wolffd@0 112 %% Action
wolffd@0 113
wolffd@0 114 % initialize
wolffd@0 115 if issparse(Ne1), Ne = full(Ne1); else Ne = Ne1; end
wolffd@0 116 clear Ne1
wolffd@0 117 [munits dummy] = size(Ne);
wolffd@0 118 Ne(find(Ne==0)) = NaN;
wolffd@0 119 for i=1:munits, Ne(i,i)=0; end
wolffd@0 120
wolffd@0 121 % Calculate neighborhood distance for each unit using reflexsivity
wolffd@0 122 % of neighborhood:
wolffd@0 123 % let N1i be the 1-neighborhood set a unit i
wolffd@0 124 % then N2i is the union of all map units, belonging to the
wolffd@0 125 % 1-neighborhood of any unit j in N1i, not already in N1i
wolffd@0 126 k=1;
wolffd@0 127 if n>1,
wolffd@0 128 fprintf(1,'Calculating neighborhood: 1 ');
wolffd@0 129 N1 = Ne;
wolffd@0 130 N1(find(N1~=1)) = 0;
wolffd@0 131 end
wolffd@0 132 while k<n & any(isnan(Ne(:))),
wolffd@0 133 k=k+1;
wolffd@0 134 fprintf(1,'%d ',k);
wolffd@0 135 for i=1:munits,
wolffd@0 136 candidates = isnan(Ne(i,:)); % units not in any neighborhood yet
wolffd@0 137 if any(candidates),
wolffd@0 138 prevneigh = find(Ne(i,:)==k-1); % neighborhood (k-1)
wolffd@0 139 N1_of_prevneigh = any(N1(prevneigh,:)); % union of their N1:s
wolffd@0 140 Nn = find(N1_of_prevneigh & candidates);
wolffd@0 141 if length(Nn), Ne(i,Nn) = k; Ne(Nn,i) = k; end
wolffd@0 142 end
wolffd@0 143 end
wolffd@0 144 end
wolffd@0 145 if n>1, fprintf(1,'\n'); end
wolffd@0 146
wolffd@0 147 % finally replace all uncalculated distance values with Inf
wolffd@0 148 Ne(find(isnan(Ne))) = Inf;
wolffd@0 149
wolffd@0 150 return;
wolffd@0 151
wolffd@0 152 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
wolffd@0 153 %% faster version?
wolffd@0 154
wolffd@0 155 l = size(Ne1,1); Ne1([0:l-1]*(l+1)+1) = 1; Ne = full(Ne1); M0 = Ne1; k = 2;
wolffd@0 156 while any(Ne(:)==0), M1=(M0*Ne1>0); Ne(find(M1-M0))=k; M0=M1; k=k+1; end
wolffd@0 157 Ne([0:l-1]*(l+1)+1) = 0;