Contents 1 EM optimization method 1.1 EM algorithm........···.·..···.· 2 l.2 Convergence..:·。。··.·····。······ 15 l.3 Usage in exponential families·.:.......·.· 19 l.4 Usage in finite normal mixtures...·..····.···· 20 1.5 Variance estimation 23 1.5.1 Louis method..... 24 1.5.2 SEM algorithm . 28 1.5.3 Bootstrap method.. 36 1.5.4 Empirical Information 37 1.6 EM Variants . 444444 38 l.6.1 Improving the E step:...·。.。·。.··· 38 1.6.2 Improving the M step 4+。·4。·。4.4。·。。。· 39 1.7 Pros and Cons.····.············· 。4 40 Previous Next First Last Back Forward 1Contents 1 EM optimization method 1 1.1 EM algorithm . . . . . . . . . . . . . . . . . . . . . . . . . . 2 1.2 Convergence . . . . . . . . . . . . . . . . . . . . . . . . . . . 15 1.3 Usage in exponential families . . . . . . . . . . . . . . . . . 19 1.4 Usage in finite normal mixtures . . . . . . . . . . . . . . . . 20 1.5 Variance estimation . . . . . . . . . . . . . . . . . . . . . . 23 1.5.1 Louis method . . . . . . . . . . . . . . . . . . . . . . 24 1.5.2 SEM algorithm . . . . . . . . . . . . . . . . . . . . . 28 1.5.3 Bootstrap method . . . . . . . . . . . . . . . . . . . 36 1.5.4 Empirical Information . . . . . . . . . . . . . . . . . 37 1.6 EM Variants . . . . . . . . . . . . . . . . . . . . . . . . . . 38 1.6.1 Improving the E step . . . . . . . . . . . . . . . . . . 38 1.6.2 Improving the M step . . . . . . . . . . . . . . . . . 39 1.7 Pros and Cons . . . . . . . . . . . . . . . . . . . . . . . . . 40 Previous Next First Last Back Forward 1