关键词: BP人工神经网络, 系统升级, bug分析, 金融软件系统

Abstract: Due to the complex business requirements, financial software system could be giant in scope and complex in system logic. During the development and maintenance of such software system, the detection and analysis of the system defects and errors are difficult and time-consuming in most of the circumstances, which usually is deemed as the bottleneck of the whole project after the midstage. In this paper, by implementing the Back Propagation (BP) algorithm, a defect analysis system was designed for the enhancement project of an Internetbased transaction trading system. The experimental results show that it does improve the developers' efficiency and productivity in defect detecting and fixing, which consequentially accelerates the project progress.

Key words: Back Propagation Artificial Neural Network (BPANN), system upgrade, bug analysis, financial software system