Generalization: EMS
Adding a kernel smoother into the EM algorithm to become EMS. The good: faster convergence by reducing dimension of parameter space; robustness; etc. the bad: introducing systematic bias. By control the bandwidth and the shape of the kernel, one may strike a balance.
PAGE NAVIGATOR(Help)
-
Controlling Infectious Diseases »Controlling Infectious Diseases
Research Questions »Research Questions
Open mathematical/computational problems? »Open mathematical/computational problems?
Statistical Methods for model selection and parameter estimation »Statistical Methods for model selection and parameter estimation
Parametric approaches »Parametric approaches
Methods incorporating both process and observation error »Methods incorporating both process and observation error
Expectation-maximization (EM) »Expectation-maximization (EM)
Generalization: EMS
+Commentaar (0)
+Citaten (0)
+About