GMM can be viewed as a soft assignment k-means, but at its core it still consists of the EM algorithm (though there are differences).
GMM is very popular and useful for clustering, especially for soft clustering, but it can also be indispensable as a tool for mimicking other datasets. It's known for being able to generate more faithful representations of minicked assets than even kmeans.
GMM can be viewed as a soft assignment k-means, but at its core it still consists of the EM algorithm (though there are differences).
GMM is very popular and useful for clustering, especially for soft clustering, but it can also be indispensable as a tool for mimicking other datasets. It's known for being able to generate more faithful representations of minicked assets than even kmeans.