Using back-propagation neural networks for functional software testing

To create reliable, safe, and high quality software is the primary purpose of software testing. With the development of software engineering, functional software testing is one essential part of software testing. This paper explains that back-propagation neural networks can be used for functional software testing efficiently and significantly. In here we established a method which is sufficient to reveal the majority of software faults in functional testing. We explained how the method can be used to produce a set of test cases covering the most common functional existing in software automatically. Then we used the proposed methodology to test large-scale software systems. Based on the results, in the paper we paid more attention to more arithmetic and date mining methods for function software testing development.