How to scale/normalize values in a matrix to be between -1 and 1

152 views (last 30 days)
I found the script that scale/normalize values in a matrix to be between 0 and 1
I = [ 1 2 3; 4 5 6]; % Some n x m matrix I that contains unscaled values.
scaledI = (I-min(I(:))) ./ (max(I(:)-min(I(:))));
min(scaledI(:)) % the min is 0
max(scaledI(:)) % the max 1
Was wondering if anyone could help me normalize values in matrix between -1 and +1 Thanks

Accepted Answer

José-Luis
José-Luis on 8 Sep 2014
Edited: José-Luis on 8 Sep 2014
Once you have your result:
scaledl = scaledl.*2 - 1;
Or directly:
result = -1 + 2.*(data - min(data))./(max(data) - min(data));
  2 Comments
Robert Hus
Robert Hus on 13 Apr 2017
Edited: Robert Hus on 13 Apr 2017
if you have an unequal spread of your data between positive and negative numbers, than the above solution may revert the sign of your array mean. To honour the original spread of positive and negative values (e.g if your smallest negative number is -20 and your largest positive number is +40) you can use the following function. Using this function the -20 will become -0.5 and the +40 will be +1. The solution above has the -20 equates to -1 and +40 to +1.
function norm_value = normalised_diff( data )
% Normalise values of an array to be between -1 and 1
% original sign of the array values is maintained.
if abs(min(data)) > max(data)
max_range_value = abs(min(data));
min_range_value = min(data);
else
max_range_value = max(data);
min_range_value = -max(data);
end
norm_value = 2 .* data ./ (max_range_value - min_range_value);
end

Sign in to comment.

More Answers (2)

Steven Lord
Steven Lord on 1 Aug 2019
If you're using release R2018a or later, use the normalize function. Specify 'range' as the method and the range to which you want the data normalized (in this case [-1, 1]) as the methodtype.
x = 5*rand(1, 10)
n = normalize(x, 'range', [-1 1])
[minValue, maxValue] = bounds(n) % Should return -1 and 1

JAY R
JAY R on 3 May 2015
Edited: JAY R on 3 May 2015
function data = normalize(d)
% the data is normalized so that max is 1, and min is 0
data = (d -repmat(min(d,[],1),size(d,1),1))*spdiags(1./(max(d,[],1)-min(d,[],1))’,0,size(d,2),size(d,2));
taken from here
  2 Comments
Steven Lord
Steven Lord on 1 Jun 2023
Suppose I told you that I had a normalized data set. Here it is.
x = [0 1];
What was the un-normalized data that was used to generate x? Any of these sets could have resulted in this normalized data.
y1 = [0 1];
y2 = [-1 1];
y3 = [42 1e6];
normalize(y1, 'range', [0 1])
ans = 1×2
0 1
normalize(y2, 'range', [0 1])
ans = 1×2
0 1
normalize(y3, 'range', [0 1])
ans = 1×2
0 1
So what additional information do you have that would let you "denormalize" x to generate a specific one of y1, y2, or y3?

Sign in to comment.

Categories

Find more on Descriptive Statistics in Help Center and File Exchange

Community Treasure Hunt

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

Start Hunting!