Simple rounding numbers in a matrix

Hello,

I have a matrix that consist in a bunch of numbers. Por exemple matrix(a)=

1.46101448605678 0.124416958027372

0.00918754579219433 -0.0655659591723838

0.549665183459168 0.218046503150596

What i want to do is make something that can round my numbers to 0, -1 and 1. If the number in the matrix is <-0,5 it should put -1 if it is bigger than 0.5 it should put 1. And the other numbers put into 0.

Thanks in advance.

Answer 1

[Azzi Abdelmalek]

A=2*rand(5)-1  % Example
A(A<-0.5)=-1;
A(A>0.5)=1;
A(-0.5<A  & A<0.5)=0

[André Pacheco]

It works, but it changes the structure of my matrix. Don't know why. My matrix is a 2x5 and after that it changes to a 1x10. Do you know why?

[André Pacheco]

Oh, i solved this initiating the new matrix with zeros.

Answer 2

[Roger Stafford]

 b = (a>.5)-(a<-.5);

Roger Stafford

Answer 3

[Image Analyst]

Andre, if you have the Image Processing Toolbox, you can quantize your array with the imquantize function in a single line of code:

m=[...
1.46101448605678 0.124416958027372
0.00918754579219433 -0.0655659591723838
0.549665183459168 0.218046503150596]

m_quantized = imquantize(m, [-0.5, 0.5])-2

In the command window:

m =
          1.46101448605678         0.124416958027372
       0.00918754579219433       -0.0655659591723838
         0.549665183459168         0.218046503150596
m_quantized =
     1     0
     0     0
     1     0

Answer 4

[Shaun VanWeelden]

Not to assume anything, but as a tutor for MATLAB, I have seen this question many, many times. If you are starting out in MATLAB, a good and easy approach is to use for loops and an if-else statement

[rows columns]=size(yourMatrix)

 for i=1:rows
  for j=1:columns
     if yourMatrix(i,j)>.5
        yourMatrix(i,j)=1
     elseif yourMatrix(i,j)<-.5
        yourMatrix(i,j)=-1
     else %its between those two
        yourMatrix(i,j)=0
     end
  end
end

I hope that helps you and anyone else, let me know if you have more questions :)

[Image Analyst]

Do your students then tend to use this loop based approach even after they've been shown the vectorized approach? Or do they switch over immediately once they learn the vectorized approach after this? I've seen people use loops for matrices that are millions of elements, even though they've been using MATLAB for months or years, when they could have used a faster vectorized approach instead.