Estimation for Binary Emax Models with Missing Responses and Bias Reduction

Documentation for package ‘ememax’ version 0.1.0

Augment_Missing	Augment missing response with observed information
Comp_Hess	Compute analytical form of Hessian matrix of Binary Emax model
Comp_I	Compute analytical form of expected information matrix of Binary Emax model
comp_Q	Maximization function estimation of EM algorithm with defined weight.
comp_Q_firth	Maximization function estimation of bias reduced EM algorithm with defined weight.
comp_theta	Estimation of emax parameters in EM algorithm iteration.
comp_theta_cox_snell	Cox–Snell bias-corrected estimator (one-step using 'clinDR' MLE)
comp_theta_firth	Estimation of emax parameters in Jeffery's prior penalized IL algorithm iteration.
comp_theta_firth_score	Firth-corrected estimating equation solution (score-based)
comp_theta_jeffrey	Jeffreys-penalized likelihood estimator via Newton–Raphson
comp_vcov	Calculate the variance covariance matrix of estimated parameters by EmaxEM
comp_vcov_firth	Calculate the variance covariance matrix of estimated parameters by EmaxEM_firth
comp_weight	Estimation of working weight in EM algorithm iteration.
fitEmaxEM	Fitting IL method with Emax model and binary response missing data.
fitEmaxEM_firth	Fitting bias reduced IL method with Emax model and binary response missing data.
log_Emax_i	Log likelihood estimation of binary Emax model
log_missing_i	Log likelihood estimation of logisitic missing indicator model
sim_data	Simulate dataset for testing Emaxem and Emaxem_firth