Fitting data to distributions

C

casta

New member
Joined
Jul 15, 2016
Messages
3
Hi guys. I am looking for a decent probability distribution to fit my data in. I have made some Q/Q-plots in SPSS. Especially the Gamma distribution does fit the data very well, and that does make sense. My data does have a large tail and has non-negative values only. However, I do not want to make my choice for a certain type of distribution on this Q/Q-plot only, since I think that this is not really 'scientific' ;)
Are there any metrics that calculate the fit of a certain model with the data?
 
casta said:
Hi guys. I am looking for a decent probability distribution to fit my data in. I have made some Q/Q-plots in SPSS. Especially the Gamma distribution does fit the data very well, and that does make sense. My data does have a large tail and has non-negative values only. However, I do not want to make my choice for a certain type of distribution on this Q/Q-plot only, since I think that this is not really 'scientific' ;)
Are there any metrics that calculate the fit of a certain model with the data?
Click to expand...
Generally speaking, obtaining the 'best probability distribution fit' to data requires iterative solutions since the parameters are not linear or even 'simple' non-linear in the parameters. Several methods can be used and are implemented in some of the statistical/math packages available such as MATLAB or language and an environment for statistical computing such as R. A nice read, IMO, is given at
https://cran.r-project.org/doc/contrib/Ricci-distributions-en.pdf
or do a search such as
https://duckduckgo.com/?t=lm&q=+Gamma+distribution+goodness+of+fit&ia=web

[url]https://duckduckgo.com/?q=distribution+goodness+of+fit&t=lm&ia=web
[/URL]
 
You must log in or register to reply here.
Top