INFERENCE OF MONTE CARLO PROPERTIES FROM THE SOLUTION OF A KNOWN PROBLEM

Abstract

A machine code was set up which was used to obtain 10,000 runs of 10 samples each from a distribution function that is a reasonable approximation to sampling distributions found in several Monte Carlo transport codes. Histograms of the probability density function for the mean value, of the standard deviation of the mean and of Student's ratio of 10 samples were obtained. Because the standard deviation has an infini e expected value, it is of doubtful usefulness in measuring the probable error of, or placing confidence limits on, the mean value of the sample. Results show that with proper precautions the sample standard deviation may be used to establish confidence li its on the sample mean in spite of the fact that it has an infinite expected value.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Jun 01, 1962
Accession Number
AD0278488

Entities

People

  • J. I. Marcum

Organizations

  • RAND Corporation

Tags

Communities of Interest

  • Air Platforms

DTIC Thesaurus Topics

  • Air Force
  • Confidence Limits
  • Data Science
  • Distribution Curves
  • Distribution Functions
  • Frequency
  • Gamma Rays
  • Gaussian Distributions
  • Government Procurement
  • Information Science
  • Machine Languages
  • Monte Carlo Method
  • Normal Distribution
  • Nuclear Radiation
  • Particles
  • Probability
  • Probability Density Functions

Readers

  • Regression Analysis.
  • Statistical inference.
  • Systems Analysis and Design

Technology Areas

  • AI & ML
  • AI & ML - Bayesian Inference
  • AI & ML - Machine Learning Algorithms