A Study on Comparative Evaluation of Software Reliability Model According to Learning Effect of Exponential-exponential Distribution
Hee-Cheul Kim · International Journal of Engineering Research and Technology · 2020
In this study, the software managers were studied the NHPP software failure model using the efficient learning factor technique for software test tools during the development and testing of software products.The life distribution using an exponential-exponential distribution was based on finite failure NHPP.The software error detection technique was composed autonomous errors-detected factor that do not know in advance that was founded automatically and learning effect, that the factors were set up by the testing administrator to detect errors in detail by prior knowledge.As a result, the model that the learning factor is greater than autonomous errors-detected factor is generally efficient model could be confirmed.In this study, the result of the mean value function pattern reflecting learning effect was show a non-decreasing form.If the model following greater the learning factor is than model following autonomous errors-detected factor, the estimated value error difference in comparison with the true value could be confirmed smaller.The overall mean square error is smaller when the learning factor is larger than the autonomous errorsdetected factor, so it can be regarded as an efficient model in terms of a measure of accuracy.Also, the larger learning factor following was the larger the estimated value of the coefficient of determination.Since all the proposed models in terms of coefficient of determination reflected on learning effect have more than 50%, all models are judged to be efficient models.So, the larger learning factor than autonomous errors-detected factor can be regarded as an efficient model in terms of goodness-of-fit.In the form of the reliability function, as the mission time passes, it gradually appears as a non-increasing pattern.And, in case of the larger learning factor than the autonomous errors-detected factor, the higher pattern shows in terms of the reliability.This study suggests that software managers can use the exponential-exponential learning factor as a basic guideline to detect software failures.