This tutorial explains how to fit a gamma distribution to a dataset in R.
Fitting a Gamma Distribution in R
Suppose you have a dataset z that was generated using the approach below:
#generate 50 random values that follow a gamma distribution with shape parameter = 3 #and shape parameter = 10 combined with some gaussian noise z <- rgamma(50, 3, 10) + rnorm(50, 0, .02) #view first 6 values head(z) [1] 0.07730 0.02495 0.12788 0.15011 0.08839 0.09941
To see how well a gamma distribution fits this dataset z, we can use the fitdistrplus package in R:
#install 'fitdistrplus' package if not already installed install.packages('fitdistrplus') #load package library(fitdistrplus)
The general syntax to use to fit a distribution using this package is:
fitdist(dataset, distr = “your distribution choice”, method = “your method of fitting the data”)
In this case, we will fit the dataset z that we generated earlier using the gamma distribution and maximum likelihood estimation approach to fitting the data:
#fit our dataset to a gamma distribution using mle fit <- fitdist(z, distr = "gamma", method = "mle") #view the summary of the fit summary(fit)
This produces the following output:
Next, we can produce some plots that show how well the gamma distribution fits the dataset using the following syntax:
#produce plots
plot(fit)
This produces the following plots:
Here is the complete code we used to fit a gamma distribution to a dataset in R:
#install 'fitdistrplus' package if not already installed install.packages('fitdistrplus') #load package library(fitdistrplus) #generate 50 random values that follow a gamma distribution with shape parameter = 3 #and shape parameter = 10 combined with some gaussian noise z <- rgamma(50, 3, 10) + rnorm(50, 0, .02) #fit our dataset to a gamma distribution using mle fit <- fitdist(z, distr = "gamma", method = "mle") #view the summary of the fit summary(fit) #produce plots to visualize the fit plot(fit)