I have a 2xN data-matrix $A$ whose rows $(x,y)$ are the number of times $y$ for which $x$ was observed (not $x$ is a number and is not categorical). If I plot $A$ using
then the result looks like a Gaussian mixture with (say) $N$-components.
How can i fit an $N$-mean Gaussian mixture? I've tried using mcclust but it does not work on my data since it is organized in a $2$-dimensional array.