Indian Journal of Science and Technology

Abstract

Objectives: In the procedure of the developing software product, the software administrators are needed to have tools/ mechanisms to detect software failures. In this research, the test implements with perspective about actual learning effects were considered based on the software reliability models. Methods/Statistical Analysis: The log-linear pattern as the lifetime distribution was offered using the finite failure NHPP. Software error detection techniques contain self-directed errors-noticed element and learning element. These two elements can be discussed from the influencing elements. When applied to log-linear property model, the learning element was more efficient than the self-directed errors-noticed element model. Findings: When applied to log-linear property model, the learning element was more efficient than the self-directed errors-noticed element model. The examination of a failure time data considering the influencing elements was achieved using software failure interval time and the maximum likelihood estimation was used for parameter estimation. Besides, the competence of the data from the trend analysis was established. And, the model choice was skilled using MSE and R2 . The reliability considering the influential elements for each model is showing a decrease pattern. In terms of reliability, the greater learning element has revealed the low reliability. In terms of judgment of the reliability, a case of the reliability for the assumed mission time, the self-directed errors-noticed element than the learning element model has shown the lower reliability. Namely, the cases of the learning element are greater or equal greater than self-directed errors-noticed element shows propensity to rise slightly, but a property of the reliability shows tend to rise progressively have the decreasing pattern for the mission time. Eventually, the reliability has sensitive property for the learning element about the mission time. Improvements/Applications: The software testing for the debugging to reduce cost in terms of the reliability from software is an essential problem. From a research, the software developers must be considered for the growth model by the prior knowledge of the software to identify failure modes which can be able to help.
Keywords: Learning Effects, Log-linear Distribution, Mission Time, Non-Homogeneous Poisson Process, Trend Analysis