conditional: Conditional Policy Evaluation

In polle: Policy Learning

Conditional Policy Evaluation

Description

conditional() is used to calculate the policy value for each group defined by a given baseline variable.

Usage

conditional(object, policy_data, baseline)

Arguments

object	Policy evaluation object created by policy_eval().
policy_data	Policy data object created by policy_data().
baseline	Character string.

Value

object of inherited class 'estimate', see lava::estimate.default. The object is a list with elements 'coef' (policy value estimate for each group) and 'IC' (influence curve estimate matrix).

Examples

library("polle")
library("data.table")
setDTthreads(1)
d <- sim_single_stage(n=2e3)
pd <- policy_data(d,
                  action = "A",
                  baseline = c("B"),
                  covariates = c("Z","L"),
                  utility = "U")

# static policy:
p <- policy_def(1)

pe <- policy_eval(pd,
                  policy = p)

# conditional value for each group defined by B
conditional(pe, pd, "B")

polle documentation built on May 29, 2024, 1:15 a.m.