0

Considering a specific PDf, how can I claim that generated samples are based on the PDF. Is it enough to compare the frequency of generated samples with the PDF curve and to show the mean of generated samples are the same as the mean of PDF?

Han
  • 23
  • Is the pdf arbitrary or are you considering a family in particular ? (normal, ...) – ippiki-ookami Jul 27 '18 at 15:12
  • You can consider the supremum between cumulative distribution functions as a metric. Comparing the means does not provide much information, since any integrable r.v. can be scaled to have the same expected. – Jo' Jul 27 '18 at 15:16
  • Many statistical tests are possible. Choosing among them depends on your purpose. What are the consequences for making an incorrect claim? – hardmath Jul 27 '18 at 15:23
  • I want to generate a network traffic based on a distribution which can be Exponential or Poisson or ... . I should show that the generated samples are based on the selected distribution. – Han Jul 30 '18 at 10:10

0 Answers0