How to normalize the values of a dataset of 32X8X2X15000 in the range -0.5 to 0.5 using standard deviation

If I have dataset of 4D single with size 32X8X2X15000. And I have calculated the mean and standard deviation of thart particular dataset. Now I want to bring all the values of the dataset within the range -0.5 to 0.5. How to do that

Answer 1

[Udit06]

Hi Anusaya,

I understand that you are trying to scale your data present in the 4-D matrix so that it falls in the range [-0.5, 0.5]. You can try out the following two approaches to scale your data:

1. Use the “rescale” function provided by MATLAB to scale the values between any given range [a,b]. You can read more about the “rescale” function using the following MathWorks documentation link: https://www.mathworks.com/help/matlab/ref/rescale.html
2. Otherwise, you can use the calculated mean and standard deviation to scale the values as shown below.

% Assuming your dataset is stored in a variable called 'data' 
data = rand(32, 8, 2, 15000) % your 4D dataset with size 32x8x2x15000 

% Calculate the mean and standard deviation 
mu = mean(data(:)); 
sigma = std(data(:)); 

% Normalize the dataset within the range -0.5 to 0.5 
normalized_data = (data - mu) / (2 * sigma); 
normalized_data = normalized_data * 0.5; 


% Clip values outside the range -0.5 to 0.5 
normalized_data = max(min(normalized_data, 0.5), -0.5); 


% Verify the range of the normalized data 
min_val = min(normalized_data(:)); 
max_val = max(normalized_data(:)); 
disp(['Minimum value: ', num2str(min_val)]); 
disp(['Maximum value: ', num2str(max_val)]); 

I hope this helps.