ABSTRACT Software Reliability is necessary to determine the wellbeing of software and find the bugs. Software reliability growth models (SRGMs) are mathematical models that interpolate software failure rates and measure the progress made during the testing. This paper suggests a supervised learning approach that uses a polynomial artificial neural network in estimating fault rate. Mean squared error (MSE) and variance are used to measure the performance of the model in real‐world data. Findings have shown that the suggested method outperforms current SRGMs, and provides better predictive accuracy and reliability values.
Sharma et al. (Sun,) studied this question.