0

I've given a dataset of N samples like: (x,y) (-100,1), (-90, 30), (-50,60), (-10,90), (0,100), (10, 90), (30, 20),(100,4)

Now I would like to determine how well the data fits a perfect gaussian (normal) distribution. For that reason, I want to fit a gaussian curve to the data and in an additional step compare the estimated gaussian with the "perfect" gaussian. Any ideas how to achieve this in C++, Boost or Eigen functions?

I'm quite stuck here. Is this even a good approach? What I actually want is a measure of how well the data follows a gaussian distribution shape.

  • Do you want to fit a gaussian curve, or determine whether the data was drawn from a gaussian probability distribution? – deftfyodor Jul 22 '16 at 17:52
  • Since I'd like to know how well it follows a perfect gaussian shape I guess it's the latter so: determine whether the data was drawn from a gaussian probability distribution –  Jul 23 '16 at 07:07
  • Many different solutions are possible. Please [search our site](http://stats.stackexchange.com/search?q=goodness+of+fit+normal+distribution). – whuber Aug 07 '16 at 20:44

0 Answers0