Bounds on Inconsistent Inferences for Sequences of Trials with Varying Probabilities

Abstract

Consider independent pairs of Bernoulli trials on two unknown sequences of probabilities p(1) = (Pi(1) : 1 < or = i < or = n) and p(2) = (pi(2) : 1 < or = i < or = n). The data are the numbers of pairs which consist of (0,0), (0,1), (1,0), and (1,1) and can be summarized in a two-way table with entries n00, n01, n10, and n11 adding up to n. The two problems of estimating the mean and variance of the number of discordant pairs n01 + n10 when Ho : p(l) = p(2) is true, and of testing Ho using the number of discordant pairs as a test statistic are considered. Two novel issues arise. While relevant parameters can not be estimated consistently from the available data, useful bounds on these can be derived. While the test is poor for alternatives typically considered in the literature, it may be effective for detecting the presence of unknown explanatory factors which discriminate between supposedly matched pairs.... Two- way tables, Matched pairs, Odds ratio, Discordant pairs.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Jun 21, 1993
Accession Number
ADA267146

Entities

People

  • Herman Chernoff
  • Yingnian Wu

Organizations

  • Harvard University

Tags

Communities of Interest

  • C4I

DTIC Thesaurus Topics

  • Boundaries
  • Classification
  • Convex Sets
  • Geometry
  • Greenhouses
  • Literature
  • Mathematics
  • Military Research
  • Observation
  • Probability
  • Random Variables
  • Security
  • Sequences
  • Statistics
  • Triangles
  • United States Government
  • Universities

Fields of Study

  • Mathematics

Readers

  • Applied Combinatorial Optimization and Logic Circuit Design.
  • Statistical inference.
  • Women's Health and Cancer Risk Research: African American Women and Pregnancy Outcomes.

Technology Areas

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