Performance Analysis of RGMs by using Supervised Learning Techniques

Software reliability is one of a number of aspects of computer software which can be taken into consideration when determining the quality of the software. Building good reliability models is one of the key problems in the field of software reliability. A good software reliability model should give good predictions of future failure behavior, compute useful quantities and be widely applicable. Software Reliability Growth Models (SRGMs) are very important for estimating and predicting software reliability. An ideal SRGM should provide consistently accurate reliability estimation and prediction across different projects. However, that there is no single such model which can obtain accurate results for different cases. The reason is that the performance of SRGMs highly depends on the assumptions on the failure behavior and the application data-sets. In other words, many models may be shown to perform well with one failure data-set, but bad with the other data-set. Thus, combining some individual SRGMs than single model is helpful to obtain more accurate estimation and prediction. SRGM parameters are estimated using the least square estimation (LSE) or Maximum Likelihood Estimation (MLE). Several combinational methods of SRGMs have been proposed to improve the reliability estimation and prediction accuracy. The AdaBoosting algorithm is one of the most popular machine learning algorithms. An AdaBoosting based Combinational Model (ACM) is used to combine the several models. The key idea of this approach is that we select several SRGMs as the weak predictors and use AdaBoosting algorithm to determine the weights of these models for obtaining the final linear combinational model. In this paper, the Fitness and Prediction of various Software Reliability Growth Models (SRGMs) can be compared with AdaBoosting based Combinational Model (ACM) with the help of Maximum likelihood estimation to estimate the model parameters.