I want to fit an appropriate distribution to my data. I have decomposed it into several levels. The first few levels start of as gamma distributed then as the levels get higher the histogram becomes more Gaussian. Hence, I thought that a generalised Gamma distribution should be a suitable fit.
So, in order to fit it, I need to estimate its (δ,α,β) parameters from my original data. Can anyone brief me with a simplified technique for accomplishing that accurately.
Thank you.
Thank you Professor Chiou for your answer. I am an Electrical Engineer, & I have just started learning statistics. Hence, I am not familiar with MLE technique. I have read about it, but haven't fully understood its approach.Would you mind elaborating more on how is it used to estimate the distribution parameters? Any hints and/or practical links would be greatly appreciated.
Thank you Professor Chiou for your clarification.
I understand now that goodness of fit is crucial in determining how good the assigned distribution fits my data. Hence, I want to try several distribution options and use goodness of fit to compare between them. One of my options is to fit my data to a mixture of distributions, for example, Gamma & Gaussian, how can I determine the contribution of each type to my data? That is, how can I determine that Gaussian contribution is .25 & Gamma is .75 for example. Is it purely empirical or is there a statistical approach?
- Badges
- Science topic