Find repeated close values in array

hello,

I have an array like

{86558,29952,29809,29804,29750,29749,29556,29382,29379,28474,15093,14974,14234,14136}

and my threshold is 3000 for checking their closeness to each other. I want to categorize them for example ~86k appears 1, ~29k appears 9, ~14k appears 3 times. I need this result like avg(~29k numbers) and their repetition time. How I can do this? I couldn't figure out. thanks.

Answer 1

[dpb]

Assume your data are sorted as shown; if not sort first.

thsh=3000;                                    % threshold for group difference screening
nGrps=sum(abs(diff(data))>thrsh)+1;           % number of groups with that difference or more
C=clusterdata(data(:),'maxclust',nGrps);      % do the cluster with those groups C is index to grp
cnts=accumarray(C,1);                         % the counts of each group
gmns=accumarray(C,data(:),[],@mean);          % compute means of each group

NB: the (:) on data array to ensure have column vector as argument irrespective of actual orientation.

Answer 2

[Image Analyst]

You can use pdist2() to find the "distance" of every element to every other element, if you have the Statistics and Machine Learning Toolbox:

v = [86558,29952,29809,29804,29750,29749,29556,29382,29379,28474,15093,14974,14234,14136]
distances = pdist2(v', v')
closeDistances = distances < 3000  % A logical matrix of whether they're close or not.