pairwise distribution function for 1D waves

Share icon

Thu, 03/19/2015 - 01:54 pm

Is there a built-in function to calculate the pairwise distribution function for points in a pair of 1D waves (especially between the same wave, i.e. an auto-pairwise distribution function)? I came up with a simple program to do so (see below), but it runs slowly (with waves that are ~10,000 points) and I expect there already is a function that does this but wasn't able to find it in the manual.

function test_distribution_function()
    wave test
    make/o/n=(numpnts(test)^2), test_DF
    variable i, j
    for(i=0;i<numpnts(test);i+=1)
        for(j=0;j<numpnts(test);j+=1)
            test_df[i*numpnts(test)+j] = test[i] - test[j]
        endfor
    endfor
   
end


Note that, despite my unfortunate naming choice, "test_DF" is not the actual distribution function. A histogram of test_DF gives the correct distribution function.
I saw the discussion in the other thread about multiplying two 1D waves to give a matrix. Here you're asking something similar (except the operation is different) and you want a 1D wave at the end. Since the matrixops are faster, this could be worth exploring to improve the speed of this procedure. Sorry I can't help more, the procedure you have is the way I'd write it.

March 19, 2015 at 03:28 pm - Permalink

MatrixOp is definitly worth exploring as that give you the best solutions.

Without it I'd write something like
Function test_distribution_function()

    Make/O/N=128 test = p
    variable numRows = numpnts(test)

    Make/O/N=(numRows, numRows) test_DF = test[p] - test[q]
    Redimension/N=(numRows * numRows) test_DF
 end

March 20, 2015 at 02:32 am - Permalink