This function estimates initial parameter values for different models using optimization techniques. The function uses the `optim` function with the L-BFGS-B method, which supports parameter bounds. It calculates the initial parameter estimates for models "M0", "M1", "M2", "M3", and "M4".
Usage
init.guess(
model,
types,
data = sim_dat$diff_Y_t,
t_list,
u_list,
init_param = rep(1, 3)
)Arguments
- model
A character string specifying the model type. Must be one of "M0", "M1", "M2", "M3", or "M4".
- types
A vector specifying the type of distribution or model for the lambda calculation.
- data
A numeric matrix with columns representing the data for each product.
- t_list
A list of vectors, where each vector contains the time or covariates for each product.
- u_list
A list of vectors, where each vector contains the additional covariates for each product (can be NULL).
- init_param
A numeric vector containing the initial parameter guesses. Default is a vector of 1's of length 3 or 5.