find combination of labels
2 Ansichten (letzte 30 Tage)
Ältere Kommentare anzeigen
elisa ewin
am 20 Apr. 2016
Bearbeitet: Jos (10584)
am 20 Apr. 2016
Hello! I describe the view of my problem: I have an user that visit same locations. This location are indicated like
A=[618 574 608];
every value in A indicate a location. To every location correspond different labels that indicate the places that a user can visit in a location:
B=[574 2;574 3;608 1;618 1;618 2;618 3;618 4];
The first column of B indicates the locations and the second column indicates the label associated to every location. E.g. Location 574 corresponds labels 2,3 Location 608 corresponds label 1 Location 618 corresponds label 1,2,3,4
What I want: starting from A and considering B, in which are indicated location and associated labels, I want to find all the possible combination of label of the sequence of location in A. The aspected output is
C=[ 121; 221; 321; 421; 131; 231; 331; 431 ]
Now I explain the output: A=[618 574 608], label associated to 618 are 1,2,3,4: the first digits of values in C; label associated to 574 are 2,3: second digits of values in C; label associated to 608 is 1: third digits of values in C.
Can you help me? I hope the question is clear: if not, I can try to explain better.
2 Kommentare
Akzeptierte Antwort
Ced
am 20 Apr. 2016
Bearbeitet: Ced
am 20 Apr. 2016
Hi
Your procedure has two steps:
1. find all labels for a current position
2. find all possible combinations
This is one possible way of doing it
A=[618 574 608];
B=[574 2;574 3;608 1;618 1;618 2;618 3;618 4];
% 1. find label for each position
N = length(A);
A_mat = cell(N,1);
for i = 1:N
A_mat{i} = B(B(:,1) == A(i),2);
end
% 2. combine
[ X, Y, Z ] = meshgrid(A_mat{:});
C = str2num([ num2str(X(:)) num2str(Y(:)) num2str(Z(:)) ]);
NOTE: I am assuming that the order of the combinations at the end does not matter? Otherwise, you will need to reorder them
EDIT: If you have more than 3 nodes in A, you can use ndgrid instead of meshgrid. Have a look at Guillaume's answer.
2 Kommentare
Guillaume
am 20 Apr. 2016
See my answer that does not depend on the size of A.
It also uses simple arithmetic to construct your final numbers instead of number to string and string to number conversions so should be much faster.
Weitere Antworten (2)
Guillaume
am 20 Apr. 2016
Bearbeitet: Guillaume
am 20 Apr. 2016
Here is one way to do it. Note that I'm using ndgrid to generate all the combinations. If you have too many locations and labels, ndgrid will run out of memory.
A=[618 574 608];
B=[574 2;574 3;608 1;618 1;618 2;618 3;618 4];
C = arrayfun(@(loc) B(ismember(B(:, 1), A), 2), A, 'UniformOutput', false); %find all labels for each location
[C{:}] = ndgrid(C{:}); %generate all possible combinations
C = cellfun(@(c) c(:), C, 'UniformOutput', false); %reshape into column vectors
C = [C{:}] %and concatenate all into one vector
%I would leave the output C as is, if you really want it as one number:
D = sum(bsxfun(@times, C, 10.^(size(C, 2)-1:-1:0)), 2)
edit: typo in the code
1 Kommentar
Ced
am 20 Apr. 2016
I believe the C = arrayfun... line should be
C = arrayfun(@(loc) B(ismember(B(:, 1), loc), 2), A, 'UniformOutput', false); %find all labels for each location
i.e. the "A" should be "loc"
Jos (10584)
am 20 Apr. 2016
Bearbeitet: Jos (10584)
am 20 Apr. 2016
A simple one-liner will do
A = [618 574 608]
B = [574 2;574 3;608 1;618 1;618 2;618 3;618 4]
C1 = arrayfun(@(k) B(B(:,1)==A(k),2), 1:numel(A), 'un', 0)
The cell C1{i} now holds the values of B(:,2) for which B(:,1) equals A(i) To get all the combinations, my function ALLCOMB (a powerful wrapper around ndgrid) makes that quite easy:
C2 = allcomb(C1{:})
ALL COMB can be downloaded from the Matlab File Exchange: http://www.mathworks.com/matlabcentral/fileexchange/10064
2 Kommentare
Guillaume
am 20 Apr. 2016
I'm unclear how that is different from my:
C = arrayfun(@(loc) B(ismember(B(:, 1), A), 2), A, 'UniformOutput', false)
and how that answer the question of generating all the combinations
Siehe auch
Community Treasure Hunt
Find the treasures in MATLAB Central and discover how the community can help you!
Start Hunting!