Software Testing Integration-Based Model (I-BM) Framework for Recognizing Measure Fault Output Accuracy Using Machine Learning Approach

In software development, the software testing phase is an important process in determining the quality level of the software. Software testing is a process of executing a program aimed at finding errors in module access, units, and involves the execution of the system being tested on a number of test inputs, and determining whether the output produced is correct. In this study, a model-based testing (MBT) called integration-based model (I-BM) framework will be developed. This I-BM framework integrates testing variables from several software testing methods, namely black-box testing, white-box testing, unit testing, system testing, and acceptance testing. The integrated variables are function, interface, structure, performance, requirement, documentation, positives, and negatives. Then, this framework will document software errors to form a dataset, which will be measured for the level of accuracy of expected manual fault output using neural network algorithm and support vector machine. From the experiment results, it shows that the accuracy level of predicting fault output values from the I-BM framework using the neural network algorithm is on average 80%, and it produces a superior SVM architecture model in predicting I-BM framework output errors with an accuracy value of 0.99, precision of 0.99, recall of 0.99, and F1-score of 0.99. Compared to other MBT, the IBM framework has the advantage of being a more comprehensive software testing model because it starts from the identification of problems, analysis, design, documentation of software testing, and recommendations for each fault output found. Thus, software errors can be classified systematically in the form of a dataset, and not only focus on software testing for product lines and module mappings.