dm.cs.tu-dortmund.de/en/mlbits/cluster-em-gmm-discussion/
Discussion of Gaussian Mixture Modeling – Lecture Notes
ISODATA process and its use in detecting compact well-separated clusters. Journal of Cybernetics . 3, 3 (1973), 32–57. DOI: 10.1080/01969727308546046
[Hofm99]
Hofmann, T. 1999. Learning the similarity of [...] Algorithm: Discussion
Complexity of Gaussian Mixture Modeling:
\(O(n \cdot k \cdot d^2 + k \cdot d^3)\) for each iteration (with a simpler diagonal \(\Sigma\) model: \(O(n \cdot k \cdot d)\) per iteration) [...] reliably estimate the parameters, we need very many data points)
matrix inversion is even \(\mathcal{O}(d^3)\)
But the general EM principle can be used with other distributions .
For example, we can use a mixture …
