1

the normal distribution plot for the data (and the rug) and the qqplot

The scatter plot for the data

I have no idea how to interpret this cullen & frey Graph

Hi, I am a student learning financial modelling. I would like some help in determining the distribution of the data given the plots above. I am reluctant to assume normal distribution of the data due to the heavy tails associated with it. The descriptive output is as follows:

summary statistics

------
min:  -6.278521   max:  6.278521 
median:  0 
mean:  -0.0001706449 
estimated sd:  0.3364102 
estimated skewness:  -1.695425 
estimated kurtosis:  250.3603 
> kurtosis(data)
[1] 246.7675
kjetil b halvorsen
  • 63,378
  • 26
  • 142
  • 467
fgkinus
  • 11
  • 2
  • 1
    Will hist, density plot, ks test give any help insights? Also, check this answer: https://stats.stackexchange.com/questions/132652/how-to-determine-which-distribution-fits-my-data-best – Sixiang.Hu Apr 10 '18 at 11:00
  • 1
    I do not seem to see the qq-plot mentioned. I would be more worried about the apparent relationship between variance and index than about normality itself. – mdewey Apr 10 '18 at 11:00
  • 1
    What is "index"? Is it simply arbitrary? What is the variable? Please show the QQ plot that you mentioned. A density plot would also be good. You can also do quantile plots against a variety of theoretical distributions. – Peter Flom Apr 10 '18 at 11:34
  • 1
    Given the super-long tails, check the [Cauchy distribution](https://en.wikipedia.org/wiki/Cauchy_distribution) – Tim Apr 10 '18 at 12:17
  • 1
    It's rarely --if ever-- the case in statistics that one can (or needs) to determine the distribution of the *data*. The objective in statistics is to determine the distribution of the *parameters*. – user603 Apr 10 '18 at 12:25

0 Answers0