Application of simple Bayesian statistics to a sample database for source correspondence

Rajesh Kumar

Research output: Contribution to journalArticlepeer-review

1 Scopus citations

Abstract

Bayesian statistics was applied to a small sample database of the tensile properties of five different brands of writing paper which were perceptibly similar. The measured parameters in the database were found to overlap for the five brands. This posed a limitation to the application of the classical approach for "match" or "no match". It was found that using Bayesian statistics for source correspondence, a mere 2-3 measurements corresponding to a particular brand raised the probabilities associated with that brand to as high a 72% and eliminating a couple of brands.

Original languageEnglish
Pages (from-to)128-131
Number of pages4
JournalForensic Science International
Volume195
Issue number1-3
DOIs
StatePublished - 25 Feb 2010
Externally publishedYes

Keywords

  • Bayesian statistics
  • Inference of source
  • Tensile strength
  • Writing paper

Fingerprint

Dive into the research topics of 'Application of simple Bayesian statistics to a sample database for source correspondence'. Together they form a unique fingerprint.

Cite this