应用神经网络技术模拟扩展可靠性数据的研究

为使可靠性评估结果有较高的准确性和可信性，要求样本量足够大，而在实际上获得这些数据是很困难的，因此能否通过模拟仿真的方式在原有数据的基础上进一步产生与原样本具有同样统计规律的数据以扩大样本量值得尝试探索，基于这样的考虑，运用人工神经网络方法对可靠性数据进行了模拟仿真分析并结合不同样本数据与传统方法进行了对比，得出了有益结论。

Study of Neural Network Simulating and Extending Reliability Data

Adequate sample sizes are necessary for reliability assessments, but obtaining reliability data is difficult. It is possible taking advantage of simulation to extend sample size which possessed of same distribution law. Beased on this idea, we apply neural network for extending reliability data and simulation statistical analysis in this paper and comparing with traditional statistical method, and arrival at a useful conclusion.