8. Statistical Analysis of Simulation Data

  1. Dirk P. Kroese1,
  2. Thomas Taimre1 and
  3. Zdravko I. Botev2

Published Online: 20 SEP 2011

DOI: 10.1002/9781118014967.ch8

Handbook of Monte Carlo Methods

Handbook of Monte Carlo Methods

How to Cite

Kroese, D. P., Taimre, T. and Botev, Z. I. (2011) Statistical Analysis of Simulation Data, in Handbook of Monte Carlo Methods, John Wiley & Sons, Inc., Hoboken, NJ, USA. doi: 10.1002/9781118014967.ch8

Author Information

  1. 1

    University of Queensland

  2. 2

    Université de Montréal

Publication History

  1. Published Online: 20 SEP 2011
  2. Published Print: 28 FEB 2011

ISBN Information

Print ISBN: 9780470177938

Online ISBN: 9781118014967

SEARCH

Keywords:

  • bootstrap method;
  • kernel density estimation;
  • Monte Carlo simulation;
  • simulation data;
  • statistical analysis;
  • steady-state performance measures

Summary

This chapter describes various statistical techniques that can be used for analyzing random data produced by Monte Carlo simulation experiments. Before embarking on a mathematical analysis of the data it may be worthwhile to detect patterns in the data through visualization and summarization. The chapter summarizes the basic estimation procedure for independent data. The procedure is sometimes referred to as crude Monte Carlo (CMC). To estimate the steady-state or equilibrium performance measure, the chapter discusses the covariance method, batch means method, and regenerative method. It also describes empirical cdf, kernel density estimation, resampling and the bootstrap method. Goodness of fit procedures can be used to assess how well simulation data fit a specified statistical model. Goodness of fit approaches fall roughly into three categories: (1) graphical procedures, (2) statistical tests based on the empirical cdf, and (3) statistical tests based on binning of the data, are described in the chapter.

Controlled Vocabulary Terms

bootstrap method; kernel density estimation; Monte Carlo methods; statistical data