Component-Based Software Testing Method Based on Deep Adversarial Network

Abstract

With the continuous updating and application of software, the current problems in software are becoming more and more serious. Aiming at this phenomenon, the application and testing methods of componentized software based on deep adversarial networks are discussed. The experiments show that: (1) some of the software has a high fusion rate, reaching an astonishing 95% adaptability. The instability and greater potential of component-based software are solved through GAN and gray evaluation. With the evaluation system, people are dispelled. Trust degree. (2) According to the data in the graph and table, the deep learning adversarial network solves the vulnerability and closedness of the general network, and the built-in test method with experimental data reaching an average accuracy rate of 90% is the best test method for this system. With the deep learning adversarial network, the average test level of component-based software reaches level 7, which makes the new software industry of component-based software have a long way to go.