【科研】如何在仿真实验中计算置信区间 - Limer98
置信区间基础详解:
1. 小白都能看懂的95%置信区间:https://blog.csdn.net/bitcarmanlee/article/details/82709774
2. 如何理解 95% 置信区间?https://www.zhihu.com/question/26419030
3. Excel 计算置信区间:CONFIDENCE函数详解:https://www.dgxue.com/excel/727.html
计算置信区间的套路:
从上面的例子来看,计算置信区间的套路如下:
1.首先明确要求解的问题。比如我们的例子,就是想通过样本来估计全国人民身高的平均值。
2.求抽样样本的平均值与标准误差(standard error)。注意标准误差与标准差(standard deviation)不一样。
3.确定需要的置信水平。比如常用的95%的置信水平,这样可以保证样本的均值会落在总体平均值2个标准差得范围内。
4.查z表,求z值。
5.计算置信区间
a = 样本均值 – z标准误差
b = 样本均值 + z标准误差
用公式表示置信区间:
论文中涉及置信区间:
[1] Oliveira, H. M. N. S. and N. L. S. da Fonseca (2018). “Protection, Routing, Modulation, Core, and Spectrum Allocation in SDM Elastic Optical Networks.” Ieee Communications Letters 22(9): 1806-1809.
[2] Wiatr, P., et al. (2015). “Energy efficiency versus reliability performance in optical backbone networks [invited].” IEEE/OSA Journal of Optical Communications and Networking 7(3): A482-A491.
[3] Lin, H. and Y. Zhuang (2019). Dynamic Multicast Traffic Grooming Using Light Trails in Elastic Optical Networks. 2019 28th Wireless and Optical Communications Conference (WOCC).
Ninety-five percent confidence intervals are computed and plotted in our simulation results.