Problem of uncertainty of graph structure identification in random variable network is considered. An approach for the construction of upper and lower confidence bounds for graph structures is developed. This approach is applied for the construction of upper and lower confidence bounds for the threshold similarity graph. The stability of confidence bounds and gaps between upper and lower confidence bounds are investigated. Theoretical results are illustrated by numerical experiments.

中文翻译：

---

随机变量网络中阈值相似度图的置信界

考虑随机变量网络中图结构识别的不确定性问题。开发了一种构建图结构的置信上限和下限的方法。该方法应用于构建阈值相似度图的置信上限和下限。研究了置信界限的稳定性以及置信上限和下限之间的差距。理论结果通过数值实验来说明。