nuisance_functions {polle}

Nuisance Functions

Description

The fitted g-functions and Q-functions are stored in an object of class "nuisance_functions". The object is a list with a fitted model object for every stage. Information on whether the full history or the state/Markov-type history is stored as an attribute ("full_history").

S3 generics

The following S3 generic functions are available for an object of class nuisance_functions:

predict

Predict the values of the g- or Q-functions based on a policy_data object.

Examples

### Two stages:
d <- sim_two_stage(5e2, seed=1)
pd <- policy_data(d,
                  action = c("A_1", "A_2"),
                  covariates = list(L = c("L_1", "L_2"),
                                    C = c("C_1", "C_2")),
                  utility = c("U_1", "U_2", "U_3"))
pd

# evaluating the static policy a=1:
pe <- policy_eval(policy_data = pd,
                  policy = policy_def(1, reuse = TRUE),
                  g_models = g_glm(),
                  q_models = q_glm())

# getting the fitted g-functions:
(g_functions <- get_g_functions(pe))

# getting the fitted Q-functions:
(q_functions <- get_q_functions(pe))

# getting the fitted values:
head(predict(g_functions, pd))
head(predict(q_functions, pd))