摘要： 针对传统软件可靠性模型需要分析软件体系结构的可行性问题，提出使用改进型神经网络计算可靠性的模型。采用自组织算法，优化设计隐含层神经网络，并使用测试数据训练网络，得到可靠性模型。实验结果证明，该模型能够在缺乏对软件内部结构分析的情况下作出与传统模型同样精确的预测。

关键词: 软件可靠性, 预测模型, BP算法

Abstract: Aiming at the problems such as feasibility to analyze the software architecture before reliability prediction, a kind of software reliability model based on improved neural network is brought forward. By using self-organizing algorithm, hidden layer optimization, and training with test data, this model is achieved. Experimental results show that the model can predict accurately as the traditional ones without the analysis of internal structure of the software.

Key words: software reliability, prediction model, BP algorithm

中图分类号: 

• TP18