在未知和一般网络干扰的实验中传递因果信息
Sadegh Shirani1, Mohsen Bayati1
1Operations, Information & Technology, Graduate School of Business, Stanford University, Stanford, CA 94305.
概括
网络干扰可能会对随机实验产生偏见. 这项研究引入了因果信息传递,以准确估计复杂网络中的治疗效应,甚至在效应稳定之前. 这种方法改善了数据驱动的决策.
科学领域:
- 统计 统计 统计 统计
- 因果推理因果推理
- 网络分析 网络分析
背景情况:
- 随机实验对于评估干预至关重要,但可以通过网络干扰使其无效.
- 网络干扰发生在一个单元的处理影响连接的单元时,偏向传统估计.
- 现有的模型与复杂和未知的网络干扰模式作斗争.
研究的目的:
- 引入一种新的框架,因果信息传递,以解决随机实验中的复杂和未知的网络干扰.
- 开发一种适用于多个单元和显著干扰的多周期实验的方法.
- 为了能够在存在网络溢出效应的情况下准确估计处理效应.
主要方法:
- 该研究的框架是基于高维的近似信息传递方法.
- 因果效应被建模为通过网络传播影响的动态过程.
- 开发了一种实用的算法来估计总治疗效果,比较所有治疗和未治疗的场景.
主要成果:
- 因果信息传递框架有效地适应复杂和未知的网络干扰.
- 该方法允许近似的潜在结果的动态随着时间的推移,提取信息之前平衡.
- 在五个数值场景中证明了有效性,具有不同的干扰结构.
结论:
- 因果信息传递为分析网络干扰随机实验提供了一个强大的方法.
- 这一框架通过超越专门的干扰模型来推进该领域.
- 开发的算法为在网络环境中估计总治疗效果提供了一个实用的工具.
相关概念视频
Interference: Path Lengths
1.3K
Consider two sources of sound, that may or may not be in phase, emitting waves at a single frequency, and consider the frequencies to be the same.
Two special sources may be considered when they are in phase. This can be easily achieved by feeding the two sources from the same source. An example would be synchronizing the two speakers by feeding them with the same source, such as the sound waves produced by a tuning fork. This setup ensures that the two sources have the same frequency and are...
Two special sources may be considered when they are in phase. This can be easily achieved by feeding the two sources from the same source. An example would be synchronizing the two speakers by feeding them with the same source, such as the sound waves produced by a tuning fork. This setup ensures that the two sources have the same frequency and are...
1.3K
Cause and Effect
10.9K
While variables are sometimes correlated because one does cause the other, it could also be that some other factor, a confounding variable, is actually causing the systematic movement in our variables of interest. For instance, as sales in ice cream increase, so does the overall rate of crime. Is it possible that indulging in your favorite flavor of ice cream could send you on a crime spree? Or, after committing crime do you think you might decide to treat yourself to a cone?
10.9K
Interference and Superposition of Waves
5.0K
When two waves of the same nature occur in the same region simultaneously, they result in interference. Interference of waves implies that the net effect of the waves is the sum of the individual waves' effects. However, it does not imply that the individual waves affect the propagation of other waves.
Interference occurs in mechanical waves, such as sound waves, waves on a string, and surface water waves. Mechanical waves correspond to the physical displacement of particles. Hence,...
Interference occurs in mechanical waves, such as sound waves, waves on a string, and surface water waves. Mechanical waves correspond to the physical displacement of particles. Hence,...
5.0K
Experimental Designs
11.1K
An experimental design is a systematic process that allows researchers to evaluate the relationship between dependent and independent variables. There are three widely used types of experimental design - pre-experimental design, true experimental design, and quasi-experimental design. In pre-experimental design, the researcher compares the data before and after some interventions or treatments. The true-experimental design has more than one purposefully created group, a commonly measured...
11.1K
Causality in Epidemiology
331
Causality or causation is a fundamental concept in epidemiology, vital for understanding the relationships between various factors and health outcomes. Despite its importance, there's no single, universally accepted definition of causality within the discipline. Drawing from a systematic review, causality in epidemiology encompasses several definitions, including production, necessary and sufficient, sufficient-component, counterfactual, and probabilistic models. Each has its strengths and...
331
Group Design
8.9K
The most basic experimental design involves two groups: the experimental group and the control group. The two groups are designed to be the same except for one difference— experimental manipulation. The experimental group gets the experimental manipulation—that is, the treatment or variable being tested—and the control group does not. Since experimental manipulation is the only difference between the experimental and control groups, we can be sure that any differences between...
8.9K


