Unbiased Monte Carlo estimation for the expected value of partial perfect information

TL;DR

This paper introduces two unbiased Monte Carlo estimators for the EVPPI, a decision-theoretic sensitivity index, improving accuracy and convergence over standard biased methods using advanced multilevel Monte Carlo techniques.

Contribution

The paper develops simple, practical unbiased estimators for EVPPI using multilevel Monte Carlo, addressing bias issues in traditional nested Monte Carlo methods.

Findings

Unbiased estimators outperform standard methods in convergence.

Numerical experiments confirm improved accuracy.

Estimators are easy to implement.

Abstract

The expected value of partial perfect information (EVPPI) denotes the value of eliminating uncertainty on a subset of unknown parameters involved in a decision model. The EVPPI can be regarded as a decision-theoretic sensitivity index, and has been widely used for identifying relatively important unknown parameters. It follows from Jensen's inequality, however, that the standard nested Monte Carlo computation of the EVPPI results in biased estimates. In this paper we introduce two unbiased Monte Carlo estimators for the EVPPI based on multilevel Monte Carlo method, introduced by Heinrich (1998) and Giles (2008), and its extension by Rhee and Glynn (2012, 2015). Our unbiased estimators are simple and straightforward to implement, and thus are of highly practical use. Numerical experiments show that even the convergence behaviors of our unbiased estimators are superior to that of the standard nested Monte Carlo estimator.