Research on software defect prediction system based on deep learning
Wei Wu, Chuyan Feng, Huimin Ren, Xiangyu Han, Xin Tong · Fifth International Conference on Mechatronics and Computer Technology Engineering (MCTE 2022) · 2022
Aiming at the high importance of software quality work in aerospace field and the requirement of high reliability of software products, we study the software defect prediction system based on deep learning; introduce the process of prediction model generation in the system from three aspects: metric selection, prediction model and evaluation indicator. The main points of design of the software defect prediction system based on TensorFlow Serving and Docker container technology are introduced from three aspects: system requirements, architecture design and model deployment. The software prediction system can be used to discover potential defects in software, improve software development and testing efficiency, and enhance the quality of aerospace software products.