IJSRP, Volume 11, Issue 6, June 2021 Edition [ISSN 2250-3153]
Qian Wang, Jian Wang
One of the research hotspots of the expectation-maximization (EM) algorithm is the parameter estimation of mixture models. Taking the Gaussian mixture model (GMM) in a multi-dimensional case as an example, we use the EM algorithm to estimate its parameters, and get the corresponding iterative formulas. Then, use R software to carry out numerical simulation experiments on the EM algorithm, and explore its convergence and sensitivity.