Counting occurrences of each column in a matrix

25 visualizaciones (últimos 30 días)
Daniel Ring
Daniel Ring el 8 de Oct. de 2018
Comentada: Daniel Ring el 8 de Oct. de 2018
I have a matrix with 16 columns and a very large number of rows. Each column has two possible outcomes in it (for example 1 or 0 in one column, .6 and .4 in another). I would like to count each of these occurrences in the columns. Preferably, I'd like to have a vector that counts each outcome of its respective column. Thank you!
[EDITED, Jan, moved from section for answers]
Example: A = [1 .6 .7 .8; 0 .4 .3 .2; 1 .6 .7 .8; 1 .6 .3 .8]
ans = [3 3 2 2 ; 1 1 2 2]
  3 comentarios
Jan
Jan el 8 de Oct. de 2018
The example is not clear: 0.8 occurs 3 times, so shouldn't the last value of the output be 3 or perhaps 1? Why is the first column of the output [3;1] and not [1;3]?
Daniel Ring
Daniel Ring el 8 de Oct. de 2018
I'm sorry, it should be ans = [3 3 2 3 ; 1 1 2 1] The first column in [3;1] because 1 occurs 3 times and 0 occurs 1 time.

Iniciar sesión para comentar.

Respuesta aceptada

ANKUR KUMAR
ANKUR KUMAR el 8 de Oct. de 2018
Since you have not given any sample data, I am taking some random data.
A=randi(50,50,50);
A(A<=25)=0.4;
A(A>25)=0.6;
A contains only 0.4 and 0.6 only, spreaded out completely. The below program gives you the occurrence of 0.4 and 0.6 in every coloumn.
nums1=arrayfun(@(x) length(find(A(:,x)==0.4)),1:size(A,2));
nums2=arrayfun(@(x) length(find(A(:,x)==0.6)),1:size(A,2));
occur=[nums1' nums2'];
nums2 can also be calculated as
nums2=size(A,1)-nums1;

Más respuestas (2)

Jan
Jan el 8 de Oct. de 2018
Editada: Jan el 8 de Oct. de 2018
Maybe:
A = [1 .6 .7 .8; ...
0 .4 .3 .2; ...
1 .6 .7 .8; ...
1 .6 .3 .8];
Count = sum(A == A(1, :), 1); % Auto-expand: >= R2016b
Result = [Count; size(A, 1) - Count];
For older Matlab versions:
Count = sum(bsxfun(@eq, A, A(1, :)), 1);

dpb
dpb el 8 de Oct. de 2018
Several ways to do this; probably easiest is via converting to categorical with the unique values in each column transformed to a single pair of categories--[0|1], [Y|N], [HI|LO], ..., whatever makes sense for the meanings.
Then histogram on those variables to return counts.
Or, use unique and the returned (optional) second index array to bin over.

Categorías

Más información sobre Resizing and Reshaping Matrices en Help Center y File Exchange.

Productos

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!

Translated by