Error Classification and Intensity Parameters in Software Reliability Models.

Abstract

The intensity parameter in software reliability models is a component of the error detection rate. This research paper explores the use of the intensity parameter in historical models. The paper shows that in the Jelinski-Moranda model the intensity parameter cannot be estimated in a meaningful manner under the usual data requirements. The paper traces the development of the models of Littlewood and of Goel and Okumoto as reactions to the Jelinski-Moranda treatment of the intensity parameter. A new software reliability model developed by Yamada and Osaki is examined. The model avoids the objections to the earlier modeling of the intensity parameter through the use of error classifications. Finally there is a discussion of research work done by Basili and Perricone in which error classification was performed. This interesting analysis is then related to the modeling work of Yamada and Osaki.

Document Details

Document Type
Technical Report
Publication Date
Sep 04, 1986
Accession Number
ADA176131

Entities

People

  • George R. Fegan

Tags

DTIC Thesaurus Topics

  • Classification
  • Intensity
  • Reliability

Readers

  • Statistical inference.
  • Systems Analysis and Design