Assessing Reliable Software using SPRT based on LPETM

Abstract

Software reliability assessment is increasingly important in developing and testing new software products. Logarithmic Poisson Execution Time Model (LPETM) is a software reliability model which predicts the expected failures and hence related reliability quantities better than existing software reliability models. It uses Non-Homogeneous Poisson Process(NHPP) with a mean value function that is dependent on exponentially falling fault detection rate. The well known sequential Probability Ratio Test(SPRT) procedure of statistical science is adopted for this model in order to decide upon the reliability / unreliability of developed software. The model is evaluated by using 6 Data Sets.

Model Prediction using Correlation In Contrast With QQ-Plot

“…The same process is repeated for all the models considered for investigation. Here we consider LPETM, HLD, and GO models for the same dataset [5][8] [9]. This process is worked out on different datasets and a sample dataset which suits all three models are considered.…”

Model Selection Using Correlation in Comparison With Qq-Plot

“…The same process is repeated for all the models considered for investigation. Here the author considers HLD, GO & LPETM Models for same dataset [5][8] [9] . The sample codeusing Rlanguage to generate the unknown parameters using MLE and further to calculate correlation factor is given below [10].…”