Using Accumarray to calculate the number of times data has fallen into a bin

Currently having a problem where I have my data variables on the LHS and the bin into which they fall on the RHS. Z is the array below with two columns. I is the number of datum (this is only a small section of Z).

45.9267 13.0000

46.1661 14.0000

46.0204 13.0000

46.0314 13.0000

45.9924 13.0000

45.9579 13.0000

45.8643 13.0000

45.4999 12.0000

45.5909 12.0000

45.4622 12.0000

45.5962 12.0000

45.8968 13.0000

45.9769 13.0000

45.7437 13.0000

45.9199 13.0000

46.1198 14.0000

46.0863 13.0000

46.3141 14.0000

46.6091 15.0000

46.3690 14.0000

46.3833 14.0000

45.7627 13.0000

46.5289 14.0000

45.5311 12.0000

47.0671 15.0000

48.1886 18.0000

The values in the right hand column are the bins in which the number in the left hand column lie.

I then have:

Z = X(1:I,2);

pieces_of_data_in_bin = accumarray(Z,1);

pieces_of_data_in_bin =

0

0

0

0

0

0

0

0

0

0

0

0

0

0

1

3

I'm not sure whether this is what I want...I basically want to have an array with [1:number of bins] on the LHS and then the number of times data has fallen into each bin on the RHS. How would I go about doing this as currently the pieces_of_data_in_bins is only showing 16 numbers, not 18 (total number of bins...).

What's going wrong?

Thanks! :)

Answer 1

[Star Strider]

You are very close. You likely need to use the unique function with ‘Z’ so the bins and their counts are correctly aligned.

Try this:

Z = X(:,2);
[Zu,~,Zx] = unique(Z);
pieces_of_data_in_bin = accumarray(Zx,1);
Out = [Zu, pieces_of_data_in_bin]

producing:

Out =
    12     5
    13    12
    14     6
    15     2
    18     1

The bin numbers are in the first column, and the counts are in the second column.

[Sophie Chumas]

Have tried it and it worked first time! But if the number of bins is big i.e. 35, then it doesn't seem to work!?

[Star Strider]

What does ‘doesn’t seem to work’ mean?

It should work unless the ‘X’ matrix has problems fitting into memory. If you are having probllems with it, please describe them and upload (attach) your data so I can work with it.

[Sophie Chumas]

This is what I'm given, and my program has to read it in:

Title = "On Road Relative Wind Velocity Magnitude and Yaw Angle"

Variables = "Time(s)", "Velocity(m/s)", "Yaw(deg)"

I=16384

0 46.933402 -2.677158

0.002 47.363864 -4.083368

0.004 47.473684 -3.907504

0.006 47.466582 -4.295546

0.008 47.394546 -4.174085

0.01 47.579053 -4.892186

0.012 47.623671 -4.529197

0.014 47.60223 -4.443917

0.016 48.016954 -4.177875

0.018 49.00

[Sophie Chumas]

I is the number of times data has been recorded I.e. the list goes on!

Thanks :)

[Star Strider]

What do you want to do with those data?

I need your data (attach the file) and a description of what you want to do with them. Is ‘I’ the row size of the data in the file?

This does not look anything at all like your original post.

[Sophie Chumas]

Ah sorry, I misunderstood! Sorry!

So the data I just posted (where there's a title and variables) is the data I'm originally given. My program is reading in the information and then I'm at the bit where I'm trying to create an array that counts from 1:number_of_bins (this is defined by the user/input function!) in one column, and then lists the number of times data has fallen into this bin (the AccumArray part). But I don't think the accumarray part is working and I'm not allowed to use the Histogram function...

Thanks! Sorry for being so confusing! :)

[Sophie Chumas]

So it seems to be showing only bins with numbers in i.e. not returning the bins that are empty!

[Star Strider]

Please also post your code so I know how you are reading your file and what you are using for variable names.

Accounting for empty bins is straightforward. In the context of my earlier code:

Out = zeros(max(Zu),2);
Out(:,1) = (1:max(Zu))';
Out(Zu,2) = pieces_of_data_in_bin

produces:

Out =
     1     0
     2     0
     3     0
     4     0
     5     0
     6     0
     7     0
     8     0
     9     0
    10     0
    11     0
    12     5
    13    12
    14     6
    15     2
    16     0
    17     0
    18     1

So at least that problem is solved.

[Sophie Chumas]

That seems to have fixed the problem - thank you very much! :)

[Star Strider]

As always, my pleasure!

[Sophie Chumas]

Thank you!

I've tried using the "contact you" page for another request but not sure if that got through!? :o

[Sophie Chumas]

Basically I'm now trying to use the Out values to define probability:

probability = (Out(:,2)/I); where I is the number of pieces of data i.e. rows...

but this produces this: which must be wrong since probability <1...

1.2222

1.2222

1.2222

3.6667

1.2222

0

0

0

0

0

0

2.4444

[Star Strider]

The denominator should not be the number of rows, but the total of ‘Out(:,2)’, so:

probability = Out(:,2)/sum(Out(:,2))

That is how I would define it.

Since it creates a vector result, if you want, you can concatenate it with ‘Out’ as either:

Out = [Out, probability];

or:

Out(:,3) = probability;

as you wish.