General Tools for Building GLM Expectation-Maximization Models


[Up] [Top]

Documentation for package ‘emax.glm’ version 0.1.2

Help Pages

emax.glm-package General linear regression via Expectation-Maximization.
AIC.em.glm Calculate the AIC of the em.glm model
BIC.em.glm Calculate the BIC of the em.glm model
data.1 Simulated data set
deviance.em.glm Model deviance (calculated from deviance residuals)
dispersion Pearson-based dispersion measurements of an 'em.glm' model.
dprob.list List of distribution functions accessed by family name ("poisson" or "binomial").
em.fit_numeric Carry our the Newton-Raphson optimization of the parameters for given weights via numeric approximations,
em.fit_pracma Carry our the Newton-Raphson optimization of the parameters for given weights via the *pracma* hessian,
em.glm Expectation Maximization glm.
em.glm_numeric_fit Numeric approximation routine
em.glm_pracma_fit Hessian routine
emax.glm General linear regression via Expectation-Maximization.
IC.em.glm General Information Criteria function
init.fit Method to initialize EM parameters. Carries out a single GLM fit and applies random noise to form starting space.
init.random Method to initialize EM parameters. Purely standard normal noise.
logLik.em.glm Calculate log-likelihood of the EM model.
make.dbinom Build a Binomial log likelihood
make.dpois Build a Poisson log likelihood
make.logLike Construct a log-likelihood function in the parameters b, for the given link family.
make_param_errors Calculate parameter errors via inversion of the Hessian matrix (either pracma or numeric approximations).
plot.em.glm Plot fit-parameters and errors
plot.em.glm.summary Error bar plot of coefficients and errors to inspect class overlap.
plot_probabilities Probability plots for the K classes fit
plot_probabilities.em.glm Test Plot em.glm
plot_probabilities.matrix Plot the class probabilities, both compared to data set index and as histogram.
predict.em.glm Predict values from an 'em.glm' model.
residuals.em.glm Deviance residuals for an 'em.glm' object.
results_k25_n1000 Simulated data set
results_k25_n1000_e05 Simulated data set
results_simple Simulated data set
select_best Select the best parameters from a set of results
sim.1 Simulated data set
sim.2 Simulated data set
sim.3 Simulated data set
small.em Carry out several short EM fits to test for optimal starting locations.
summary.em.glm Summarize EM glm coefficients.
summary.small.em Summarize a small.em class
update_probabilities Construct normalized class properties for a given set of parameters