Averaging points within data set

Hi,

I have some data that I have got from an engine. This data is pressure from a certain point with the engine. To acquire this data I had the engine complete 200 revolutions of the full 720 degree cycle.

My issue now is that i have 200 sets of 720 data points.

The output data goes;

data set 1 - degree 1

data set 1 - degree 2

data set 1 - degree 3

...all the way to degree 720.

data set 2 - degree 1

data set 2 - degree 2

data set 2 - degree 3

...all the way to degree 720.

data set 3 - degree 1

data set 3 - degree 2

data set 3 - degree 3

...all the way to degree 720.

All the way to data set 200

Is there a way that I can set the average value of degree 1 from data set 1, data set 2, data set 3...all the way to data set 200 and be able to do this for all of 720 degrees?

[dpb]

Well, are the data all in one file, 200 files or in memory as a very long vector or what, precisely? It'll be simple enough to do once we have a more precise definition of just what a "data set" really is...

[Lewis Parry]

The 200 sets of data are within a single txt file. I have put an image of the data I get in the post below

Answer 1

[Andrei Bobrov]

x=importdata('yourfile.txt');
xavg=accumarray(rem((0:numel(x)-1).',720)+1,x(:),[],@mean);

[Andrei Bobrov]

I am fixed my answer.

Answer 2

[Mike Caldwell]

Do you mean you want to GET the average across all of your data sets? If performance isn't an issue, you could use for loops. Assuming your dataset is in matrix form (if not you could arrange it like this):

Arrangement:
   row 1: 1 ... 720
   row 2: 1 ... 720
   . . . . . . . .
   row 200: 1 ... 720

Code:
  average = zeros(1, 720);
  sum = 0;
  for i = 1:720
    for j = 1:200
        sum = sum + dataset(j, i);  //add all the values in ith column
    end
    average(i) = sum/200;   //then take the average
    sum = 0;
  end

Hope that answers it. -Mike

[Lewis Parry]

Hi Mike thank you very much for your help. What you have said is helpful, but I don't think that I explained the way that the data was set out very well.

The data come in a single column text file with it giving the data for dataset1 - sample 1 sample2 sample3 ... sample 720 dataset2 sample 1 sample2 sample3 ... sample 720.

I have attached an image of the data that I get and an excel document of the data I want to get from the file.

Thanks for your help!

Answer 3

[dpb]

I can't read the image to tell what the data file name is, but given the data are just a linear sequence so don't have to do any file manipulations to get them in a useful sequence the solution is almost trivial--

x=importdata('yourfile.txt');    % read all the data in the column
xavg=mean(reshape(x,720,[]),2);  % average the 200 columns of 720 rows by row

What this relies on is internal storage order and using the vector abilities in Matlab to arrange the data into an array by the sequential samples. You have 720 samples for each cycle so rearranging by that as the first dimension leaves the total number/720 columns, for each of the "tests". Then, use the optional second DIM argument to the mean function to average over those columns. (Of course, you could also transpose the result of the reshape operation and average columns and get the same result.)