Parameter Estimation of Software Reliability Growth Model using Metropolis-Hastings Algorithm and Application of Sequential Probability Ratio Test

Purpose: This study employs the Metropolis-Hastings algorithm to determine the software reliability growth model (SRGM) parameters and introduces the sequential probability ratio test (SPRT) process.BRMethods: The actual failure dataset was fitted to the Yamada imperfect debugging model 1, and the parameters were determined using the M-H algorithm. In SPRT, the interval scale was used in 15 cases.BRResults: We estimated the parameters for the Yamada imperfect debugging model 1 using the M-H algorithm. In addition, we discussed SPRT in 15 cases . Consequently, we established the reliability of the dataset early in the data collection process.BRConclusion: The M-H algorithm can be used instead of the maximum likelihood or Bayesian methods to estimate the SRGM parameters. Although we used a gamma distribution in this study, future investigations can consider other distributions. In addition, real development environments can benefit economically from using SPRT.