Problem in finding PDF of a set of samples

2 次查看（过去 30 天）
may
may 2013-9-30
评论： Image Analyst 2013-10-1
I want to get PDF of a set of samples, A, from its histogram:
[n,x] =hist(A,Number_bins);
when I use the following code:
pdf=n/sum(n(:));
with different number of bins I get different results, to fix it I use this code
pdf=n/sum(n(:))/diff(x(1:2));
but some of numbers in pdf would be greater than one! (the probability should not be greater than one)
I really don't understand where the problem is. I would appreciate if you you could help me. Thank you.

请先登录，再回答此问题。

回答（1 个）

Image Analyst
Image Analyst 2013-9-30
Of course the PDF is different if you have different bin sizes. But most important is how you plan on using the PDF. What do you plan on doing with it? I think n/sum(n) is fine - you just have to know what you're dealing with when it comes time to use it. It tells you the probability of a sample falling into that bin.
  3 个评论
显示 1更早的评论
may
may 2013-9-30
编辑：may 2013-9-30
I edited the question, thanks.
Image Analyst
Image Analyst 2013-10-1
Don't divide by the diff. And of course the PDF will vary according to the number of bins. Computers are digital - you can't have an infinite number of bins. It can't be continuous. It must be quantized and since the number of counts in the quantized pdf must equal the number of elements in your array, the value per bin must be different. My other question, about what you plan on doing with it, remains unanswered.

请先登录，再进行评论。

类别

在 Help Center 和 File Exchange 中查找有关 Histograms 的更多信息

标签

产品

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!

Translated by