What distribution is shown by this histogram?

  • Thread starter Thread starter Nyasha
  • Start date Start date
  • Tags Tags
    Distribution
AI Thread Summary
The discussion centers on a histogram generated from 100,000 datasets of 80 random Poisson distributed numbers, with p-values calculated for each dataset. Participants inquire about the specific algorithm or transformation used to derive the histogram and express curiosity about techniques for identifying data distributions. While goodness-of-fit tests are acknowledged, the conversation emphasizes the need for methods to determine the underlying distribution before analysis. It is suggested that if a specific probability distribution is assumed, the empirical distribution should reflect a "probability of a probability." The focus remains on understanding the distribution represented by the histogram.
Nyasha
Messages
127
Reaction score
0
So l generated a 100 000 data sets in Matlab each consisting of 80 random poisson distributed numbers. I then calculated the p-values for each data set and stored them into a vector which l used to make this histogram. I was wondering, what distribution is shown by this histogram
 

Attachments

  • what distribution is this.JPG
    what distribution is this.JPG
    18.8 KB · Views: 400
Physics news on Phys.org
Hey Nyasha.

Can you outline what algorithm or transformation of the 80 Poisson distributed simulator calculated results you did?
 
Just curious: what type of general techniques there are to recognize the distribution

of data? I know there are ways of comparing against a normal,but not of a way of

deciding the general distribution of some data. And yes, I know there are

goodness-of-fit tests, but this is more an after-the-fact approach.
 
if Nyasha can calculate a "p-value" then obviously he must have assumed some specific probability distribution for some statistic. If his assumption is correct, the empirical distribution he had should approximate a "probability of a probability". For example if a random variable X is normally distributed then one may define another random variable Y as: Y = the probability of the value of X that was observed.
 
Back
Top