2

there.

I always appreciate the members belong this site because very active support!

Now I have some data set which were measured by same experimental setup, but their distributions were slightly different I think. Distribution of some look like Gaussian distribution, the other look like Laplace distribution.

Let me attach some examples. enter image description here

I have two questions as below.

1) How can I estimate its distribution as Laplace or Gaussian?
 What I did is just from my decision by their shape.
 Though I searched some threads from here, there seemed no answer about that.

2) What is difference between two distributions?
 I measured same stuffs(but different models) with same experimental setup.
 I thought there distribution were same but actual cases were different.

Thank you in advance!

TZakrevskiy
  • 22,980
actlee
  • 376

1 Answers1

2

I am not sure to have fully understood the questions. Maybe what you are looking for are statistical tests on distributions. If you want to test that your data follow a given distribution you can use the Kolmogorov–Smirnov test. With this test you will be able to compare the theoretical quantiles of the Laplace distribution (for example) to your observed ones.

If you want to simply test the normality of the distribution, the Jarque–Bera test can be useful and is based on the skewness and Kurtosis.

user66444
  • 121