Causal Discovery from a Mixture of Experimental and Observational Data

因果知识构成了我们在科学中所知道和想知道的很多东西。这项研究的目标是开发和应用一种统一的方法，从观察和实验数据的混合中表示和发现因果关系。概率因果网络正被用作因果关系的一种表示。贝叶斯方法正被应用于从数据中学习概率因果网络。这种发现方法的计算机实现正在使用由人类专家构建的现有因果模型产生的数据进行经验性研究。混合使用观测和实验数据的能力将极大地扩展贝叶斯因果建模和发现在科学中的应用范围。Http://www.cbmi.upmc.edu/~gfc/index.html/causal_discovery

Causal knowledge makes up much of what we know and want to know in science. The goal of this research is to develop and apply a unified method for representing and discovering causal relationships from a mixture of observational and experimental data. Probabilistic causal networks are being used as a representation of causality. Bayesian methods are being applied to learn probabilistic causal networks from data. A computer implementation of the discovery method is being investigated empirically using data generated from existing causal models that were constructed by human experts. The ability to use a mixture of observational and experimental data will expand considerably the scope of application of Bayesian causal modeling and discovery in science. http://www.cbmi.upmc.edu/~gfc/index.html/causal_discovery