Updating alpha in estimation procedure. Updating alpha by fitting a weighted multinomial regression.
Arguments
- alpha
a matrix of alpha before updating - regression coefficient for multinomial logistic regression model
- beta
a matrix of beta before updating - class specific parameters for recurrent model
- d
a vector of observed recurrent events for subjects of interest
- Z
a vector of time-independent corvariates
- mu_censor
a vector of estimated mu(C), where C is a vector of censoring time
- gamma
individual frailty. 0 represents the frailty equals 1 and k represents the frailty follows gamma(k,k)