Finding frequency of values in an array

Hi there

I have a 128x128x16 double array. Firstly, I would like to get the total number of values < zero in this array. Then, i would like to get the sum of the frequency of occurance of each value in this array. The array name is Phase.

thanks

Answer 1

[Star Strider]

One approach:

negPhase = nnz(Phase < 0);                          % Number Of Values < 0
[uPhase,ia,ic] = unique(Phase);
tally = accumarray(ic, 1);
freqOccurrence = [uPhase, tally];                   % Frequency Of Occurrenct Of Each Value

If you do not want the results sorted (since the unique function does this by default), use:

[uPhase,ia,ic] = unique(Phase, 'stable');

instead. The rest of the code is unchanged. (Note that this will conform to the MATLAB linear matrix indexing convention, so it will be a column vector, not a matrix.)

[Jan]

I tried it with histcounts, but was frustrated about the change from histc, which still confuse me. n = histcounts(Phase, uPhase) replies numel(uPhase)-1 counts. This is documented, but not useful. I know, that accumarray does the trick, but I still hope, that I have overseen how histcounts can count the elements directly, like the name seems to imply.

[Ahmad Alenezi]

Thanks for your reply.

what does this psection mean:

[uPhase,ia,ic] = unique(Phase);

And, how can i sum all frequencies (only) as numbers ?

i used below script to sum the frequencies, but did not work and gives me (6831x2) double array:

sumfreq = sum(freqOccurrence, 2);

[Star Strider]

This unique call:

[uPhase,ia,ic] = unique(Phase);

gets the unique values in ‘Phase’ and returns a sorted vector of them and index vector ‘ic’ for use with the accumarray call. It is necessary because of the way accumarray works, and returns its output.

Your assignment here:

sumfreq = sum(freqOccurrence, 2);

calculates the sum of the individual rows of ‘freqOccurrence’. If you want to sum the second column of ‘freqOccurrence’, do this:

sumfreq = sum(freqOccurrence(:, 2));

That should do what you want.

[Leyla Elyasizad]

Hey Guys

Thanks for your helpful comments but I tried both unique & tabulate for double values with 4 decimal places (0.0202, 0.0031,.....) and with both functions I get two 0.0031 with differenct occurence!

Do you have any idea why this happens?

[Dyuman Joshi]

@Leyla Elyasizad, because both numbers do not appear to be the same. See below -

%Assign 0.3
y = 0.3

y = 0.3000

%Value st   
sprintf('%0.42f', y)

ans = '0.299999999999999988897769753748434595763683'

Read more here - FAQ: Why is 0.3 - 0.2 - 0.1 not equal to zero?

When trying to compare floating point numbers, it is better to use a tolerance.

y = 0.12230455

y = 0.1223

sprintf('%0.42f', y)

ans = '0.122304549999999997988986422114976448938251'

%Set tolerance of 10^-6
tol = 1e-6;
%%Comparison
%Using equality
isequal(y,0.12330455)

ans = logical
   0

%Using a tolerance
abs(y-0.12230455)<tol

ans = logical
   1

With the same idea in mind, try uniquetol instead of unique

Answer 2

[dipanka tanu]

can you use tabulate operator of matlab ?

[Star Strider]

Try it!