Random values from histogram
1 Ansicht (letzte 30 Tage)
Ältere Kommentare anzeigen
Hi
how i can use one value at the time from histogram and after use that value is deleted and i cannot use that again eg. if i have
x=rand(1,100)
[N,X]=hist(x,15)
bar(X,N,1,'w')
i want to get 100 values one by one
Thanks a lot
2 Kommentare
Akzeptierte Antwort
Jos (10584)
am 11 Feb. 2011
Standing on the shoulder of a giant WR, the following run-length based code will create a random list of 10 values that are derived from the results of a histogram
% simplified example
% data obtained by hist
X = [.1 .4 .6 .8] ; % possible values
N = [5 3 0 2] ; % how many times the occurred
% a simple but effective run-length decoding scheme
clear ix ;
ix([1 cumsum(N(1:end))+1]) = 1 ;
ix = cumsum(ix(1:end-1)) ;
% retrieve the values
X2 = X(N>0) ;
V = X2(ix) ; % all 10 values
RandV = V(randperm(numel(V))) % in random order
0 Kommentare
Weitere Antworten (2)
Walter Roberson
am 8 Feb. 2011
Okay, then, what this is equivalent to is taking each bar count as a run-length encoding of the value associated with that bar (e.g., the bar center), then doing a randperm() of that expanded vector.
I know a couple of people have posted efficient run-length decoding routines; I just don't happen to remember one at the moment.
0 Kommentare
Walter Roberson
am 8 Feb. 2011
x(randperm(1:length(x))
The histogram becomes irrelevant. Each item will occur exactly the number of times it did in the original list.
3 Kommentare
Walter Roberson
am 8 Feb. 2011
Hmmm... the bar _counts_ are what are to be chosen from, perhaps? But if so then should the bar count decrease one one of the items from that bar is "used up", or should the bar count itself stay the same but the number of repetitions of it "available" decreases?
Or perhaps it is the bar _index_ that needs to be chosen randomly, in accordance with the counts?
Siehe auch
Kategorien
Mehr zu Matrix Indexing finden Sie in Help Center und File Exchange
Produkte
Community Treasure Hunt
Find the treasures in MATLAB Central and discover how the community can help you!
Start Hunting!