Search code examples
rnormal-distributionmultivariate-testingiris-dataset

How to fit a multivariate normal distribution in R?

I need to fit a multivaraite normal distribution to each specie in the Iris dataset in R. I saw the mvtnormpackage might be useful; however, i want to use the maximum likelihood estimation and not sure how to do so in R. Any ideas?

Solution

  • It seems you are looking for a mixture discriminant analysis (since class labels are known). In this case, you can use the MclustDA from the mclust package.

     model= MclustDA(data = iris[,1:4], class = iris$Species)
 summary(model)

    However, if you are looking to cluster the data using by fitting multivariate Gaussian mixtures then you can use the Mclust function.

      fit = Mclust(data = iris[,1:4], G=3)
  table(fit$classification, iris$Species)

 #   setosa versicolor virginica
 # 1     50          0         0
 # 2      0         45         0
 # 3      0          5        50