This function calculates the empirical value functions as defined in equation (8) in Section 2.2 of the manuscript.

evf(data, pred, propensity)

Arguments

data

a data frame containing the ID (ID), outcome (reward), and observed treatment (treatment) information of subjects.

pred

a matrix with K(K-1)/2 columns, where K is the total number of treatments. Each column is the recommendations between two treatments for all subjects.

propensity

a data frame with K columns, where K is the total number of treatments. The kth column is the propensity scores of assigning the kth treatment to subjects.

Value

A scalar as defined in equation (8) in Section 2.2 of the manuscript.