I am new to GMM training. I could see from the link as below
The main limitation of the GMM algorithm is that, for computational reasons, it can fail to work if the dimensionality of the problem is too high (i.e. greater than 6 dimensions for instance).
Could you anyone please provide a reference/justification for the claim. I am planning to have GMM trainer for a maximum of 16 features. I would like to know in which scenarios this will happen.