Make a matrix smaller based on some criteria

6 views (last 30 days)
I am working with a matrix.
A=[...
20 40 21.5 41
5 21 15 10]
Now, first of all, I want to find out the similar values from row 1. Lets say, the similar values are 20 and 21.5, another pair is 40 and 41. For the each pair, I want to eliminate one column from each pair based on their corresponding column value in the second row. So, I want to eliminate
41
10
and
20
5
because its value in the second row is smaller. The final matrix that I want is
21.5 40
15 21
Thanks in Advance.

Accepted Answer

Md
Md on 2 Jan 2014
row 1 is 20 40 21.5 41 and second row is 5 21 15 10

More Answers (1)

Image Analyst
Image Analyst on 2 Jan 2014
You'll need to specify how close is close enough to be removed. Is it 1 away? 0.5 away? 2 away? Then I'd start scanning columns, and if any column is too close to another one, delete it:
% Delete column 42:
A(:,42) = [];
  3 Comments
Image Analyst
Image Analyst on 2 Jan 2014
OK, I think the algorithm should work. Just make sure you use a while instead of a for because the array will be shrinking as your scan along and you need to take that into account. Post your code if you have trouble.
Image Analyst
Image Analyst on 2 Jan 2014
Edited: Image Analyst on 2 Jan 2014
See if this works for you:
rows = 2;
columns = 40;
A = randi(99, rows, columns) % 2 by 40 matrix of values from 1-99.
tooClose = 2;
column = 1;
lastColumn = columns;
while column < lastColumn
% Find difference between current column and all the other columns.
differences = abs(A(1,:) - A(1, column));
% Find out which ones are too close and need to be deleted.
columnsToDelete = find(differences <= tooClose);
if length(columnsToDelete) > 1
% Don't delete the last one.
columnsToDelete = columnsToDelete(1:end-1);
% Now delete those columns
A(:, columnsToDelete) = [];
% Get new size.
[rows, lastColumn] = size(A);
end
% Increment column we're looking at
column = column + 1;
end
% Report final version of A to the command window.
A

Sign in to comment.

Categories

Find more on Matrices and Arrays in Help Center and File Exchange

Tags

Community Treasure Hunt

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

Start Hunting!