A function to compute the log-likelihood of the COM-Poisson model.
comp_mu_loglik(param, y, xx, offset, summax)
comp_mu_neg_loglik_log_nu_only(log_nu, mu, y, summax)
numeric vector: the model coefficients & the current value of nu
.
It is assumed that nu
is in the last position of param
.
numeric vector: response variable
numeric matrix: the explanatory variables
numeric vector: a vector of length equal to the number of cases
maximum number of terms to be considered in the truncated sum
numeric: nu in log-scale
numeric vector: fitted mean parameters
comp_mu_loglik
returns the log-likelihood value of the COM-Poisson model based on Huang (2017).
comp_mu_neg_loglik_log_nu_only
returns the negative log-likelihood value of the COM-Poisson model based on Ribeiro Jr et al. (2018)'s specification to use in conjunction with optim
.