Bootstrapping Interval Estimations for Software Reliability Assessment Based on a Discrete NHPP Model  
Author Shinji Inoue


Co-Author(s) Hironori Matsuno; Shigeru Yamada


Abstract We discuss a bootstrap method for software reliability assessment based on a discretized nonhomogeneous Poisson process (NHPP) model. Ordinarily, model parameters of the discretized NHPP model are estimated by using the regression analysis based on the regression equation derived from a difference equation of the discretized NHPP model. However, it is not so easy to derive some information for the statistical inference on software reliability assessment by the existing estimation approach for the model parameters because it is very difficult to identify a probability distribution function for the parameter estimates analytically. In this paper, we discuss a method for estimating probability distributions and confidence intervals for the model parameters based on the discretized NHPP model by applying a bootstrap method to the regression analysis, and show numerical examples of five types of the bootstrap confidence intervals for the several software reliability assessment measures by using actual data.


Keywords Software reliability growth model, Bootstrap confidence interval, Software reliability assessment, Nonhomogeneous Poisson process model
    Article #:  1954
Proceedings of the 19th ISSAT International Conference on Reliability and Quality in Design
August 5-7, 2013 - Honolulu, Hawaii, U.S.A.